Research Briefing: Understanding Trends in Hate Crimes Against Immigrants and Hispanic-Americans

NIJ was directed by Congress to evaluate trends in hate crimes against immigrants and Hispanic-Americans and to assess the underlying causes behind an increase in hate crimes against such groups.

In October 2010, NIJ contracted with Abt Associates to analyze trends in hate crimes using multiple data sources in selected states. On May 10, 2011, the researchers — Jack McDevitt, Professor, Northeastern University; Michael Shively, Senior Associate, Abt Associates; and Rajen Subramanian, Associate, Abt Associates — presented a briefing on their work so far.

Summary of the Research Briefing

The project is divided into two phases:

Phase I features:

Phase II, to begin in August 2011, involves fieldwork and surveys in the four selected states intended to provide contextual and supplementary information to augment and interpret the foundational secondary analyses conducted in Phase I.

The presentation provided an overview of the project, the state selection process and preliminary findings of the Phase I analyses. The focus was on the assessment and analysis of data from the UCR and assessments of the utility of the NIBRS and NCVS data in pursuing the key research objectives.

State Selection Process

The objectives of the selection process were to select four states with:

  • Numbers of hate crime cases in key databases sufficient to support statistical modeling.
  • Solid hate crime laws and statistical reporting processes and accessibility of data.
  • Geographic and demographic variation.
  • Active engagement of key sectors addressing hate crime: social services, advocacy groups and criminal justice.

Objective indicators weighed in the selection process include data on:

  • The percent of the population who are foreign-born (Census).
  • The percent of the population of Hispanic descent (Census).
  • Rates of change from 2000-2009 in foreign-born and Hispanic populations (Census).
  • Estimated number of illegal immigrants (U.S. Citizenship and Immigration Services field report, 2000).
  • The number of agencies submitting reported hate/bias crime incidents (UCR).
  • Whether state bias crime laws include race/ethnicity.
  • The number of agencies with ICE 287g Memoranda of Understanding (MOAs). A set of qualitative indicators was also considered, and the end result was the selection of California, Michigan, New Jersey and Texas. The study will also examine Arizona in a supplemental analysis in Phase II.

Preliminary Findings of Secondary Data Analyses

The UCR data are the most promising for analyzing trends and have the potential to be used for program evaluations, particularly data with a national scope and structured to allow time series designs. The data provide enough statistical power and sufficient reporting and coverage of the U.S. to determine the statistical significance of substantial events in a time series, such as the increase in reported hate crime following the attacks of September 11.

The models of UCR data also suggest that there may be a slight rise in hate crimes against Hispanics from 2004-2008, although the effect is modest when some modeling techniques are used, and disappears in other types of models.

Figure 1: Modeled Trends for All Hate Crime, UCR Data

Figure 2: Modeled Trends for All Anti-Hispanic Hate Crime, UCR Data

Annual Counts of Hate Crimes Captured by the UCR, 1996-2008
Year Total Hate Crimes Anti-Hispanic
Hate Crimes
1996 8759 566
1997 8048 491
1998 7751 483
1999 7875 466
2000 8063 559
2001 9730 599
2002 7462 481
2003 7489 426
2004 7649 476
2005 7163 522
2006 7722 576
2007 7624 595
2008 7783 562

The table "Annual Counts of Hate Crimes Captured by the UCR, 1996-2008" presents the raw counts of all hate crimes and anti-Hispanic hate crimes in the UCR data for each year 1996-2008. The counts were discussed in order to show the challenge of analyzing and modeling trends below the national level, given the limited number of events captured by the UCR each year. When examining the hate crime data in the table and comparing them to the trends presented in the graphs, it is important to keep in mind that the graphs represent models, rather than graphic presentation of the counts.

Other preliminary findings of Phase I assessments of the data and analyses discussed in the presentation include:

  • State-level UCR data can detect significant trends in the two states with the greatest number of events (California and Texas), but not in other states.
  • The utility of NIBRS hate crime data is compromised by incomplete coverage across states. The data are best used to assess the traits and composition of hate crime incidents and not trends in the number of incidents. The NCVS data allow the assessment of national trends, but cannot be used at the state level.

The presentation concluded with a discussion of Phase II activities and how the information acquired over the next year would be used to confirm and augment what was learned through the secondary analyses of Phase I.

Date Modified: May 25, 2011