Introduction

There is little question that the America’s racial/ethnic demographic profile is transforming rapidly. The changing complexion of American society is influenced by numerous factors, chief among them is the exponential growth in the size of the Latino population. Currently, nearly one-sixth of the United States population is Latino, a level which is expected to rise to nearly one-quarter in the coming decades.Footnote 1 In the span of 10 years (2000–2010), the Latino population increased by nearly 25%, a level triple that for the entire population (Owens 2010). To place this increase into perspective, Owens (2010) points out that the growth in the Latino population alone accounted for roughly half of the country’s population increase between 2000 and 2006 and much of dramatic rise is attributable to immigration, documented, and undocumented alike.Footnote 2

The Latino rise for both internal migrants and immigrants has affected settlement patterns. Latinos remain highly concentrated within the five states (California, Texas, Florida, New York, and Illinois) that have historically served as destination points.Footnote 3 In all likelihood, these states will continue to draw and house the majority of the Latino population into the foreseeable future. Yet, the size of the Latino population has surged in a number of non-tradition settlement states. For example, since 2000, the Latino population grew by more than half in a number of Southern states typically not known for a strong Latino presence—Arkansas (61%), Georgia (59%), South Carolina (57%), Tennessee (56%), and North Carolina (55%).Footnote 4 This population shift is due to a number of factors, including the combination of economic/social opportunities among these states and changes in migration patterns due to aggressive law enforcement security/apprehension along the US-Mexico border (Nevins 2002).

Although scholarly attention is paid to Latino demographic trends, less is known about the extent to which the recent rise in the Latino population may be associated with patterns of criminal justice or violent criminal outcomes. This is not to suggest that criminologists are not sensitive to, or aware of, the ongoing changes or the need to incorporate Latinos in scholarly examinations of these issues (see Martinez 2002; Martinez 2006; Lee and Martinez 2009). Instead, we submit that given the dynamic nature of the Latino growth in this country, it is important to continue to update current knowledge on this subject; that is, to provide timely information on a subject that is the subject of ongoing public and political discourse.

The current study has this goal in mind. The primary objective of this research is to provide a contemporary overview of the Latino experiences with the criminal justice system as well as the impact of Latino concentration on levels of lethal violence. First, we highlight ethnic/racial disparities in incarceration and sentencing. Next, we present results from both regional (i.e., focus on Border States) and national analyses. The results from both sets of multivariate analyses offer insight into the overall association between Latino concentration and homicide victimization generally and for an area of the country that is characterized as particularly vulnerable to violent crime. Quantitative examinations at either level, though substantively valuable, have been the subject of an increasing number of research projects. We conclude with a discussion of the continued relevance of this subject for future public policy considerations as well as directions for future research.

Latinos in the Criminal Justice System

Similar to African Americans, over the years, Latinos (especially immigrants) are believed to have a higher propensity to engage in criminal violence. Cohen (1996) and Santana and Smith (2001) found that Latinos were over-represented in local media violent crime stories in San Antonio, San Francisco, and Florida, respectively. In 2000, the General Social Survey published by the National Opinion Research Center showed 73% of Americans believed immigrants were either “somewhat” or “very” likely to increase crime (Rumbaut 2008). Later in 2006, after the immigrant crime debate gained momentum, US Representative Steve King (R-Iowa) asserted a day without immigrants would save the lives of 12 US citizens who would otherwise die a violent death at the hands of murderous illegal aliens each day (Beirich 2007). Although immigration is often cited as a crime facilitating social process, we submit that this notion is often extended to Latinos more generally because often US citizens (of non-Latino origin) do not know the nativity status of Latinos. In other words, unlike obvious outward physical racial appearances that have long been markers of stereotypes, often the general public conflates ethnicity or ancestry with place of birth.

This broad generalization takes on increased relevance in light of the fact that Latinos are becoming increasingly geographically diversified. As Latinos settle in areas where they have traditionally accounted for a small share of the population, it is likely that their minority status will become more visible. That is, unlike in the historical destination areas, the growing presence of Latinos in communities in Southern states, for example, may translate into their over representation in the criminal justice system, owing to discrimination or other institutional practices (Harris et al. 2009; see also Steffensmeier and Demuth 2000).

More broadly, we argue that questions remain as to whether one result of the growth of the Latino population nationally and among the newer destination states is an increased involvement with law enforcement institutions. Racial/ethnic disparities in criminal justice outcomes are most commonly presented as comparisons between non-Latino whites and non-Latino blacks, largely leaving open the question of how Latinos fare as they become involved with various institutions of social control. We argue that it is important to document the extent to which, as a product of contact with the broader criminal justice system, the experiences of Latinos resonate with those of their African American counterparts (see Harris et al. 2009). Below, we turn to a review of recent federal and state-level trends in courts and corrections.

In terms of incarceration trends over the last decade, one important pattern, according to Mauer and King (2007: 2), is the “growing proportion of the Hispanic population entering prisons and jails.” More specifically, in federal, state, and local prison/jails, Latino inmates increased from 16% in 2000 to 20% in 2008 (West and Sabol 2009), yet the share of Latinos in the US adult population rose from 11 to 13% over the same period (Lopez and Livingston 2009). The preponderance of evidence suggests that the United States is witnessing a rise in federal incarceration rates, and a disproportionate number of Latinos are being incarcerated across state prison/jails when compared to their black/white counterparts. Continued exposure of Latinos to the criminal justice system, at these alarming rates, belies simplistic views that Latinos are more crime prone and may suggest more systemic flaws within the US criminal justice system that directly and indirectly promotes higher rates of incarceration for minorities. Similar to the experience of blacks, Latino interaction with police, courts, and corrections is producing low confidence in, and high exposure with, the criminal justice system (Lopez and Livingston 2009).

Federal

A report issued by the Pew Hispanic Center in 2009 found that increased enforcement of immigration laws (e.g., illegal crossings) has altered the ethnic Latino composition of offenders sentenced in federal courts (Lopez and Light 2009). Media outlets have also reported that increased border enforcement, tougher laws, and recent speedy large-scale prosecutions (e.g., use of guilty-pleas) under Bush’s Operation Streamline has contributed to the sharp uptick of Latino incarcerated offenders (e.g., Moore 2009). Lopez and Light (2009) reported that in 2007, four-in-ten (40%) offenders sentenced in federal courts were Latinos, compared to whites (27%) and blacks (23%), respectively. In terms of geography, more than half (56%) of all Latino offenders were sentenced within five of the nation’s 94 US district courts. Unsurprisingly, the courts are positioned along the US-Mexico border: Southern (17%) and Western (15%) districts of Texas, the District of Arizona (11%), the Southern District California (6%), and the District of New Mexico (6%). What is more, of the Latinos sentenced during this year, approximately seven-in-ten (70%) did not have US citizenship.

Two border security and enforcement programs have been linked to the rise of Latino federal offenders. The first was President Clinton’s 1994 Operation Gate Keeper initiative, which employed human resources and sophisticated technology (e.g., underground sensors and face recognition) aimed at restoring security and safety along the nation’s busiest border crossings. The second was President Bush’s 2005 Operation Streamline program. Originally launched in Del Rio, Texas, this program advocated a zero-tolerance policy by convicting immigrants charged with illegal border crossings. The program has been characterized as controversial because the process involves taken immigrants en masse (e.g., 66 defendants) and fast-tracking through the federal system. For example, from inception to court adjudication, immigration cases spanned 2 days when compared to 460 days for white-collar and 333 days for drug prosecutions (Schwartz 2009). In 2007, among Latino offenders, 37% were sentenced for drug offenses, while 48% were sentenced for immigration offenses (Lopez and Light 2009). Latinos who did not have US citizenship were sentenced to 40 months, compared to US citizen Latinos who received longer sentences (61 months). Immigration offenders received the shortest prison sentences, averaging 25 months.

Critics argue that federal resources have been diverted to immigration cases instead of egregious violent/drug offenses, and more importantly the use of coerced pleas. Prior to Operation Streamline, illegal crossing offenders were rarely charged with federal misdemeanors and transported back across the border. In US v. Roblero-Solis, the US Court of Appeals held that Operation Streamline’s en mass plea hearings in Tucson, Arizona, was in direct violation of federal law, in that, proceedings were not sufficiently “personal.” An article published in the New York Times states that “[t]he immigration system has essentially become criminalized at a huge cost to the criminal justice system, to courts, to judges, to prisons, and prosecutors” (Moore 2009). The convergence of criminal law and immigration has been described as a process known as “crimmigration” (Stumpf 2006). The implications for the rise in Latino federal prisoners is that the system will undergo a substantial increase in poor Latino (and poor black) representation, a prison culture with limited English proficiency, and a need for services dedicated to such growth (Moore 2009). Given its novelty, little is known how other federal programs such as 287(g), which permits state, county, city officers, and employees to perform the functions of federal immigration officers, might further agitate the Latino experience.

States

While a 2010 Prison Count study showed that for the first time in 40 years, the number of states prisoners in the United States has declined (The Pew Center on the States 2010), Latino incarceration disparities remain a problem for state jails/prisons. The Sentencing Project study showed that nationwide, Latinos (1.8) were incarcerated nearly double that of whites (Mauer and King 2007). Put differently, for every 100,000 persons incarcerated, 742 are Latino, compared to 412 whites. Mauer and King (2007) also projected that 1-in-6 Latino and 1-in-3 black males born today can expect to serve some time in prison/jail. Lopez and Livingston (2009) reported that in 2007, approximately 4% of adult Latinos were either in prison, jail, probation, or parole, compared to whites (2%) and blacks (9%). Mauer and King (2007) also identified states with disparate Latino incarceration rates. For example, Connecticut, Massachusetts, and Pennsylvania showed Latino to white incarceration ratios more than twice the national average. States with greater concentrations of Latinos showed incarceration disparities below the national average: California, Texas, Florida, and Nevada. Using more recent 2008 data, BJS reported that Latino incarceration rates remain disproportionately higher than whites, yet lower than blacks (Sabol et al. 2009).

Conclusion

Given the broader state drop in jail/prison population due to residential-/community-based diversion programs, increased parole, and shortened probation terms, there is a possibility that Latino rates of incarceration might also trend downward over the next few years (Prison Count 2010). A more likely circumstance that may contribute to a decrease in future Latino federal incarceration rates (especially among non-US citizens apprehended for border crossings) is the observation from multiple sources, suggesting that Mexican migration to the United States has been reduced to nil (Passel and Cohn 2010; Wolgin and Garcia 2011; Cave 2011; Stevenson 2011). Fueled by an older Mexican population, drop in birth rates, improved economic, and educational opportunity, Mexicans are opting to remain in their country. Other factors that have deterred the migration north into the United States is the poor American economy, risk of apprehension (due to increased border security), and higher fees ($700 in 1980s to $2000 in 1990s) charged by human traffickers to smuggle across the border. For the first time in 60 years, net traffic has been characterized as zero to negative (Cave 2011). The Pew Hispanic Center reported that the average inflow of undocumented immigrants from Mexico fell from around 500,000 per year from 2000 to 2005 to only 150,000 per year in 2007–2009 (Passel and Cohn 2010).

Despite the observed increase in the size of the Latino population that has contact with the criminal justice system, we believe that it is somewhat premature to conclude that this is suggestive of the fact that Latinos have a higher propensity to engage in criminal violence. Certainly such a conclusion is consistent with the popular perceptions and the common rhetoric used by anti-immigration advocates and pundits calling for increased levels of border security as a means of reducing crime (see Nevins 2002; Martinez et al. 2008). Put another way, it may be tempting for some to seize upon these facts as clear evidence that as the Latino population continues to grow, and particularly due to the arrival of immigrants, crime rates in this country are also likely to increase. However, anecdotal evidence does not lend credibility to such claims, as it has been well established that national violent crime rates have been on the decline during the same period as the rapid expansion of the Latino population (see Blumstein and Wallman 2000; Stowell et al. 2009). In the discussion to follow, we take up the issue of the nature of the association between Latinos and criminal violence.

Latinos and Criminal Violence

To this point, our discussion has focused broadly on Latino representation, or involvement, with the criminal justice system. A related issue is the manner in which Latino concentration shapes patterns of violence. To address this question, we will examine models that concentrate first on counties in states along the US-Mexico border and then we will present results from national county-level analyses. This approach will allow us to extend current knowledge on two important areas of inquiry, which have not yet been thoroughly examined in the criminological literature.

The border analyses are central to studies on this subject for at least two reasons. First, the areas adjacent to the border are often a focal concern of policy makers since they are thought to be exceptionally violent and affected most directly by the negative consequences of unauthorized immigration (see Orrenius and Coronado 2005). Scholars have observed that it is not uncommon for state and local governments to push for more stringent border security as a means of reducing local crime (see Nevins 2002; Martinez et al. 2008). Second, despite the strong public concern regarding crime and unauthorized immigrants, the evidence is again at odds with the perceptions. In both individual-level and macro-level studies, researchers fail to find support for the notion that undocumented migrants “pose a unique threat to public safety” (Hickman and Suttrop 2008, 77; see also Hagan and Palloni 1999). While these findings are suggestive, few macro-level criminological studies have focused explicitly on this region. Similarly, quantifying the effect of Latino concentration on levels of violence at the national level is similarly useful, as it will allow us to offer information about how this relationship may operate differentially beyond the areas that fall along the border. The key contribution in these analyses will be the identification of whether and to what extent Latino concentration affects homicide levels in the areas that have experienced the sharpest increases in Latinos, those often referred to as “new” Latino destinations. Before reporting the results, we will provide a brief summary of the data and analytical method used for the multivariate models.

Data and Methods

The homicide information was drawn from the county-level data compiled by the Centers for Disease Control (CDC). Again, to ensure the stability of our estimates, the homicide data were pooled for the 5-year period from 1997 to 2001 (see Martinez et al. 2008; Lee et al. 2001). These data are particularly well suited for the current study because they provide information on the race/ethnicity of victims. The ability to disaggregate these data racially and ethnically allows for a comparison of the potential differential impacts of compositional or social structural factors on levels of lethal violence, a level of specificity which cannot be achieved using data sources traditionally employed in macro-level criminological research (i.e., Uniform Crime Report).

In the current study, our choice to focus exclusively on homicide victimization is guided by both data availability limitations and concerns regarding the reliability of dependent variables constructed for expressions of non-lethal violence. In terms of the former, at present, there are no publically available national-, county-level data sources that provide information for victims or offenders based on Hispanic/Latino origin. Furthermore, we recognize that although homicide is a rare event, it remains less sensitive to under-reporting as compared to other expressions of non-lethal violent and property crimes (e.g., assault, robbery and larceny). To the extent that there are reporting and classification differences for such criminal outcomes both across various law enforcement agencies from different states and between agencies within states, this may generate unreliable estimates. For example, one possibility is that law enforcement practices may be less formal in rural county jurisdictions, thereby introducing a degree of non-comparability to the resulting crime measures (Osgood and Chambers 2000).

However, homicide is less subject to such uncertainties because it is more clearly identifiable, and thus, different law enforcement agencies nested within different counties from different states are likely to have established procedures that yield a higher degree of uniformity in reporting. Given these reliability concerns and because the CDC data allow for a more refined examination of the differences in factors that predict victimization across racial/ethnic groups, we believe that our concentration on lethal violence is warranted.Footnote 5

Demographic and social structural information for the independent variables was drawn from Census 2000 files and includes an established array of indicators used in previous research on this topic. We constructed the array of independent variables in a way that is consistent with previous criminological research. Socioeconomic disadvantage is a composite measure that is the sum of z-scored values for five social structural dimensions—poverty, unemployment, female-headed households with children, non-Latino black composition, and public assistance receipt. Residential stability is a measure that includes both the percent of the population that have moved within 5 years and the percentage of housing units that are vacant. Our measure of immigrant concentration is constructed using the percent of the population who are foreign-born and the share of the population who are Latino. In addition, we include a number of other control variables; the percent of the civilian labor force employed in professional; the ratio of adults to children; and the percent of the population who are men between 18 and 34 years (for a broader discussion of variable definitions see Martinez et al. 2008; Stowell 2007). Due to the highly skewed nature of the homicide distributions, we estimate the regression models using Poisson-based negative binomial regression techniques (see Osgood 2000).Footnote 6

Results

US-Mexico Border

As mentioned above, areas adjacent to the US-Mexico border are often portrayed as being particularly susceptible to violent crime. Indeed, these areas are often cited as dangerous, “lawless” communities, owing to the presence of undocumented immigrants, a segment of the Latino population generally considered to be disproportionately criminally inclined. For example, concerns over the increased levels of criminal violence and the compromised public safety caused by illegal immigration were cited as some of the principle considerations underpinning Arizona’s controversial legislation SB1070. In her official statement regarding the law, Arizona’s Governor Jan Brewer indicated that she became convinced to sign the bill into law, in part, due to her concerns about “border-related violence and crime due to illegal immigration.”Footnote 7 Such sentiment remains widely held, although very little in the way of empirical evidence is offered in support of such claims.

In an initial examination of this contention, we compare the overall and racially/ethnically specific homicide rates for counties along the border. The rates of lethal violence in these counties are compared to the rates in the non-border counties in the States of Arizona, California, New Mexico, and Texas (see Fig. 1). In this figure, it is clear that total homicide rates in the border counties mirror those for the region, which raises questions regarding the claims which characterize the border as uniquely violent and dangerous. This pattern holds even when we disaggregate by the race/ethnicity of the victims. With the exception of the black victimization rates, there is a relatively high degree of consistency between the rates of lethal violence independent of the proximity to the border. Furthermore, we observe that Latino homicide rates are actually lower along the border, though only marginally, than in non-border counties and the region more generally. Based on the descriptive findings, there is no compelling evidence to suggest that communities near the border are anomalous in terms of their levels of lethal violence.

Fig. 1
figure 1

Comparison of homicide victimization rates in border and non-border counties, by race and ethnicity. Source: Homicide data provided by the Centers for Disease Control (years 1997–2001)

Full size image

The results from the negative binomial regression analyses for counties within the four Border States are presented in Table 1. Focusing first on the total homicide levels (column 1), we observe that a number of social structural factors (socioeconomic disadvantage, ratio of adults to children) emerge as predictors of levels of lethal violence, a pattern that has been documented in previous studies on this topic (see Martinez et al. 2008; Peterson and Krivo 2005). More central for the purposes of the current study is the fact that, net of other structural and demographic characteristics, counties adjacent to the border have significantly lower levels of lethal violence than non-border counties. Further, immigrant concentration (an indicator that captures both the size of the Latino population and the share of the total population that is foreign-born) is unrelated to homicide levels. The null effect of immigrant concentration indicates that, counter to popular perceptions, the arrival (and presence) of Latinos is not associated with higher levels of lethal violence (see Lee and Martinez 2009; Sampson 2008; Stowell 2007).

Table 1 Negative binomial regressions for homicides by race and ethnicity, counties in Border States (N = 292)
Full size table

In the race-/ethnic-specific models, the border effect is relatively robust, also predicting lower levels of homicide for non-Latino blacks, Latinos, and approaches statistical significance for non-Latino whites. Again, this result runs counter to the conventional wisdom that communities along the border as distinctly more violent, due to the social disruption caused by undocumented immigration. Interestingly, the results in Table 1 also underscore the value of running racially/ethnically disaggregated models, and the importance of community context as it relates to how structural conditions shape the likelihood of victimization differently across groups. Specifically, while for each group structural factors such as economic disadvantage and residential stability are systematically associated with homicide, as expected by theory and previous research on this subject, the impacts are not uniform. For example, disadvantage is a stronger predictor of non-Latino black homicide victimization than it is for non-Latino whites or Latinos. Residential stability, however, is found to be positively related to homicide victimization risks for these two groups, while inversely associated with homicide levels for non-Latino blacks. Consistent with prior research, we also find that immigrant concentration is associated with lower levels of non-Latino white and black homicide victimization, and we observe a null effect for Latinos.

National

Given the changing geographic settlement patterns of Latinos, especially over the past decade, it is important for criminologists to consider the impact of Latino concentration/immigration and violence more broadly. In the following discussion, we present the results from a series of national county-level regression models. We believe that it is important to research the question from a national perspective in an effort to bolster our understanding of how patterns of lethal violence are influenced by the presence of Latinos. This is a salient line of inquiry given the widespread opposition to immigration and the rapidly changing compositional characteristics within new Latino destination areas (see Archibald and Thee-Brenan 2010; Unz 2010). Specifically, in the subsequent regression models, we include an indicator of hyper-Latino growth, a designation identifying states in which the Latino population doubled between 1990 and 2000.Footnote 8

Figure 2 includes comparison of the trends in overall homicide rates for the nation and between hyper-growth and non-hyper-growth states. It is clear from this figure that homicide patterns within the hyper-growth states, though higher than the national average, followed the well-documented declines in lethal violence over the past decade (see Blumstein and Wallman 2000). The elevated rates of homicide victimization among the hyper-growth states are not entirely unexpected, given that rates of violent crime have historically been higher among Southern states, owing in part to a phenomenon researchers have attributed to a “culture of violence” (see Baller et al. 2001; Ousey and Lee 2010). Furthermore, despite the exponential growth in the size of their Latino populations, the levels of homicide in these states did not increase.

Fig. 2
figure 2

Comparison of trends in state-level homicide rates (1994–2004) by increase in size of Latino population. Source: Homicide data provided by the Centers for Disease Control (years 1997–2001)

Full size image

The evidence indicates that levels of lethal violence in the hyper-growth states dropped more sharply than they did in non-hyper-growth regions or the nation as a whole.Footnote 9 Between 1994 and 2004, the homicide rate fell by nearly 32% in the states that experienced the largest increases in the size of their Latino population, compared to drops of 25 and 27% for the nation and non-hyper-growth states, respectively. We do not want to overstate the importance of these differences, as all of the reductions are of a similar magnitude. Still, we argue that the descriptive findings are instructive because they do not reveal, as conventional wisdom would hold, that the arrival of Latinos (either via immigration or secondary migration) is a disruptive and crime-generating social process (see Lee and Martinez 2009; Martinez et al. 2010) (Fig. 3).

Fig. 3
figure 3

Percent change in state-level homicide rates (1994–2004) by increase in size of Latino population. Source: Homicide data provided by the Centers for Disease Control (years 1997–2001)

Full size image

The results from the regression analysis are presented in Table 2. In these models, we again observe that economic disadvantage and occupational characteristics of counties are systematically associated with levels of homicide, a finding that is consistent with expectations. In addition, in three of the four models, immigrant concentration predicts lower levels of homicide, which is supportive of prior research on this topic (see Lee and Martinez 2009; Peterson and Krivo 2005; Sampson 2008). That is, the combined effect of the size of the Latino population and the percent foreign-born is associated with fewer total, non-Hispanic black, and Latino homicides. The immigrant concentration coefficient in the non-Latino white model is positive and null. Taken together, the combination of negative and null effects presented here offers further support for the notion that immigrant concentration is associated with lower levels of lethal violence (Lee and Martinez 2009; Sampson 2008).

Table 2 Negative binomial regressions for hyper-Latino growth and homicides by race and ethnicity, US counties (N = 336)
Full size table

Finally, the hyper-Latino growth indicator is statistically significant for total and non-Latino white homicide models. The null effect in the racially/ethnically disaggregated models suggest that group-specific homicide levels in the states that experienced the largest increases in Latinos do not differ significantly from the other states. This significant effect for total and non-Latino white homicides is not entirely surprising, as research has shown that levels of lethal violence tend to be higher among Southern states, where a large share of the nation’s white population, and by extension white homicides, are concentrated (see Baller et al. 2001). Considering these facts together, it seems that the areas into which Latinos are settling may have historically higher levels of homicide and not that the expansion in this segment of the population (due to the combination of immigration, internal migration, and fertility) is necessarily promoting a higher risk of lethal violence. Again, we believe that it will be crucial for criminologists to continue to monitor levels of violence in these areas over time before firm conclusions can be offered regarding the impact that the changing Latino composition may hold for criminal violent outcomes, or that Latinos in particular may have greater violent criminal tendencies.

Conclusions

Latinos remain the largest and fastest growing segment of the population and there is an increasing number of few criminological studies in which they are the primary research subject. Increasingly, scholars are moving beyond the standard practice of focusing on disparities between non-Latino whites and blacks in terms of criminal justice outcomes and criminal involvement (Lee and Martinez 2009; Martinez 2002, 2006; Martinez et al. 2008; Sampson 2008). The present study was designed to bring current knowledge and empirical evidence to bear in these two conceptually distinct, but overlapping, research areas. Because of the centrality of border-related violence in discussions calling for increased public safety, we highlighted this region in our regional and national-level regression analyses.

The results from our review of Latino contact with the criminal justice system do suggest that over the last decade, this segment of the population is increasingly represented in the all levels of the incarcerated population. Definitive explanations for this increase are not readily available at present, though we caution against a summary conclusion that this is a reflection of an increased propensity to engage in violent crime. It is not yet clear how much of this increase is attributable to systemic practices such as discrimination or harsher sentencing for Latinos (see Steffensmeir et al. 2009). In our estimation, this topic represents a fruitful line of inquiry for future research.

Furthermore, our quantitative analyses do not lend support to the argument that Latino concentration is a socially destabilizing, and ultimately a crime facilitating process. As support, we offer results from regional analyses, which indicate that counties in states situated along the US-Mexico border tend to have lower levels of homicide than the remaining counties in those states. Our regional findings also fail to show strong evidence that immigration/Latino concentration is positively associated with levels of lethal violence, as many lawmakers, political pundits, and much of the general public contend. The results from the national-level regressions similarly reveal that the concentration of immigrants/Latinos is associated with lower levels of homicide victimization generally, and for each of the racial/ethnic groups. Finally, when controlling for counties within states that experienced the sharpest increases in the size of their Latino populations, these counties tended to be characterized by higher levels of lethal violence generally, yet this effect did not emerge in the racially/ethnically disaggregated models. In short, the findings from our analyses are not consistent with the notion that increased violence is a natural consequence of the general or geographic increases in the Latino population.

We believe that the results from the present study cast new light on issues relating to Latinos, crime, and their experiences with the criminal justice system. That said, we also recognize a number of limitations in the present study, which would have allowed for further exploration on this subject. We encourage future researchers to undertake, as the results from such studies will advance current knowledge regarding the Latino experiences within American society. Ideally, we would have included a wider array of criminal outcomes, both violent and property-based, in order to test for differential racial/ethnic effects on various crime types. Similarly, macro-level research that considers the race/ethnicity of offenders will contribute to a more comprehensive understanding of the social structural factors that predict criminal involvement by race. Such data are difficult to collect and compile, and as noted above, may be subjected to unreliability based on reporting practices (see Cancino et al. 2009 for exception). Preliminary research that examines various criminal outcomes, though not racially/ethnically disaggregated, does not offer a strong reason to anticipate results that vary widely from those presented here, yet such variations have yet to be quantified in the research literature (see Stowell 2007; Stowell and Martinez 2007; Stowell et al. 2009).

Data availability also prevented us from examining this question using a longitudinal analytical approach. We recognize that research concentrating on the dynamic nature of the demographic shifts among Latinos and criminal justice outcomes will provide important information for researchers and policy makers. Understanding the impact of rapid compositional and social structural changes has historically been, and continues to be, of great interest to social scientists. A number of longitudinal studies examining the temporal impact of immigration have been conducted; however, we believe that there is also an opportunity to focus scholarly attention on Latinos more broadly (Martinez et al. 2010; Ousey and Kubrin 2009; Stowell et al. 2009).

The aforementioned limitations notwithstanding, we argue that the current study is able to provide new insight into the contemporary Latino experience in the United States. More importantly, as the size and diversity of the Latino population continues to grow in this country, the issue of their incorporation into American society will likely continue to hold the attention of academics and policy makers. Recent estimates indicate that over the next several decades, one-quarter of the United States population will be of Latino descent (Owens 2010). With the influx of Latino immigrants and internal migration into new communities, it is likely that this will be met with ongoing concern regarding the impact their presence has on levels of violent criminal deviance. We encourage future researchers to conduct studies on this topic, as there is indeed much important work yet to be undertaken to maintain current knowledge on this dynamic social process (see Martinez 2002). Now, and in the coming years, studies that focus on the experience of Latinos will certainly have implications for theoretical development, enhance criminological understandings of the processes underlying the relationship between demographic/nativity changes and crime, and offer sound empirical evidence for the purposes of policy making.