Top Ten Violent Crime Counties in the United States


Top Ten Violent Crime Counties in the United States

Based on Department of Justice Victimization Data

Violent Crime Rates-2010-2012. Published in December, 2015

  1. Sacramento County, CA—44.1
  2. Allegany County, PA—42.4
  3. Philadelphia County, PA—41.4
  4. Honolulu County, HI—39.4
  5. Pima County, AZ—36.1
  6. Dallas County, TX–35.9
  7. Riverside County, CA—34.4
  8. Alameda County, CA—33.5
  9. Marion County, IN—30.6
  10. Salt Lake County, UT—29.7 and Milwaukee County, WI—29.7

Editor’s note: Considering that the above estimates are based on victimization surveys conducted by the US Department of Justice, the numbers are larger than those using crimes reported to law enforcement through the FBI’s Uniform Crime Reports. Thus, the rankings will be more stable/accurate than those based on FBI data.

See for a list of the top ten most violent metro areas.



This is the first time that the Department of Justice, using the National Crime Victimization Survey (NCVS), have offered comparison data of states, counties and metropolitan areas (not cities).

The data within this report that will be of interest to most readers is contained on pages 44-47 where crime rates (2010-2012) are offered for counties and metropolitan areas.

This report will, undoubtedly, set off an endless series of comparisons that will be new and unique to local and state crime discussions.

It may prove to be one of the most discussed crime-related documents in the country.

News editors will ask why one jurisdiction does so well when compared to others. For example, California’s Sacramento County has a violent crime rate of 44.1 percent and Orange County has a violent crime rate of 11.1 percent. Why are the differences so large? Are there policy implications? What will the administrators of Sacramento County have to say?

Many readers who are used to reports based on crimes reported to police via states and eventually to the FBI (Uniform Crime Reports or UCR) will be confused by this report, which is based on a national survey of crime victims.

Comparing the NCVS to the FBI’s UCR data

The FBI’s UCR data have historically been the primary source for state-and substate-level crime statistics. The UCR estimates are based on crimes as reported by participating law enforcement agencies, although the FBI partially adjusts the state estimates for missing data from some jurisdictions.

State-level UCR estimates are used in this report for purposes of comparison to the SAE (Small Area Estimates) estimates. Differences between national estimates from the NCVS and the UCR have been noted for decades, and a substantial body of research has attempted to account for these differences.

At the national level, the NCVS and UCR trends in crime were not always consistent during the period 1999–2013, and the NCVS shows a larger decline in violent crime rates across the period compared to the UCR.

Several differences between the NCVS and the UCR could account for the discrepancies, including the following:

The UCR is based on police reports, while the NCVS asks survey respondents about crimes regardless of whether they reported the crimes to police.

The UCR covers crimes committed against all persons and businesses, while the NCVS covers persons age 12 and older and excludes crimes against the homeless, institutionalized populations, and businesses.

The two collections differ in regard to the types of crimes included. In the UCR, violent crime includes murder and nonnegligent manslaughter, forcible rape, robbery, and aggravated assault.

In the NCVS, violent crime includes rape and sexual assault, robbery, aggravated assault, and simple assault. Simple assault, which is not included in the UCR, is the largest component of the NCVS estimate of total violent crime.

 From the Report 

The National Crime Victimization Survey (NCVS) is the only source of annual national data on a number of policy relevant subjects related to criminal victimization, including intimate partner violence, injury from victimization, weapon use, the cost of crime, reporting to police, and crime against vulnerable populations, such as the elderly and persons with disabilities.

It is one of the two main sources of data on crime in the United States and the only source that provides detailed information on the level, nature and consequences of crime (The Nation’s Two Crime Measures).

By capturing crimes not reported to police, known as the “dark figure of crime,” as well as those known to law enforcement, the NCVS serves as the primary, independent source of information on crime in the United States.

Though the NCVS was originally designed to provide national-level estimates of criminal victimization, the Department of Justice through the Bureau of Justice Statistics (BJS) has recognized an increasing need for victimization data at the state and local level.

Three major reviews of the NCVS program all point to the demand that local criminal justice administrators have for empirical information to shape policy. Subnational estimates are of value to both federal and nonfederal data users and stakeholders.

Federal stakeholders that currently allocate funding or resources for crime victims and crime prevention based on official police crime estimates, could use the victimization estimates to understand how the allocation of funding would change when unreported crime is taken into account.

Policy makers could use these estimates to examine local variations in crime both reported and unreported to police and make comparisons among states, and law enforcement officials could use the findings to begin to understanding differences in rates of crime and reporting to police across. The data can also be used in conjunction with official police statistics to begin to understand the correlation between the NCVS and official police reports of crime.

Research demonstrated that the NCVS could be enhanced to produce several types of subnational estimates. The small area estimates (SAE) presented in this current report use statistical methods to generate model-based subnational estimates for all 50 states, DC and other large places.

In addition to these model-based estimates, BJS is developing other approaches for generating subnational NCVS estimates. BJS is boosting the NCVS core sample in large states to obtain direct state-level estimates. Beginning in July of 2013, BJS initiated a pilot test boost of the NCVS sample in the 11 states with the largest population to assess the feasibility and cost of generating direct state-level estimates. Based on findings from the pilot test, BJS is boosting the NCVS core sample in 22 states and largest metropolitan areas.

Next, BJS has started developing generic area typologies based on various geographic, social, economic, or demographic characteristics. These generic areas represent all places that are similar to each other based on the characteristics of interest.

Initially these subnational geographic identifiers include region, population size, and urbanicity.

Finally, BJS is testing the feasibility of fielding a low-cost mailed, self-administered companion survey in specific cities. The American Crime Survey research project is focused on developing and evaluating a cost-effective supplement to the core NCVS that will help BJS and law enforcement better understand the counts and rates of crime at the local level.

Collectively, the BJS will assemble and evaluate the various approaches based on estimate quality, relevance, level of geography, timeliness, burden, and cost to develop a routine approach to generating annual subnational estimates. The SAE estimates presented in this report will be labeled as developmental until the BJS implements the methodologies that will allow it to produce these estimates on an annual basis in conjunction with estimates from the other components of the subnational estimation program.

Although the estimates in this report are labeled as “developmental” the methodology behind them has gone through rigorous review and verification.





  1. […] Source: […]

  2. […] We previously published a list of the ten most violent counties (using the same methodology) at […]

%d bloggers like this: