Is American criminal justice color-blind? The statistics say no

By John Roman :: July 16th, 2013


Noting that there are racial disparities in the American criminal justice system is hardly newsworthy. From stop-and-frisk to motor vehicle searches at traffic stops to sentencing, racial disparities abound in modern America. However, until the George Zimmerman trial for Trayvon Martin’s death, one gaping disparity had received little attention.

Black Americans are far less likely to be adjudicated as justified in using deadly force in a firearm-related death. The difference between rates of justifiable rulings in cases with a white shooter and a black victim and cases with a black shooter and a white victim are astonishing.

In fact, they dwarf every other racial disparity in an already racially unbalanced criminal justice system. The differences are so great that any notion that justice in America is color-blind is at risk.

The numbers below require little explanation. Drawing from Supplemental Homicide Reports (SHR) submitted by local law enforcement to the FBI between 2005 and 2010, we see that in cases with a white shooter and a white victim, the shooting is ruled to be justified less than 2 percent of the time. If the shooter is black and the victim is white, the rate of justifiable homicide rulings drops to almost 1 percent. However, if the shooter is white and the victim is black, it is ruled justified in 9.5 percent of cases in non-Stand Your Ground (SYG) states. In SYG states, the rate is even higher—almost 17 percent.

Now consider the situation that occurred in the Zimmerman case (and I note that none of these facts are in dispute). When there is a homicide with one shooter and one victim who are strangers, neither is law enforcement, and a firearm is used to kill, a little less than 3 percent of black-on-white homicides are ruled to be justified. When the races are reversed, the percentage of cases that are ruled to be justified climbs to more than 29 percent in non-SYG states and almost 36 percent in SYG states.

The one gap in the SHR data is the setting where the homicide occurred. If it turns out that almost all the white-on-black homicides occur in residences or businesses and almost all the black-on-white homicides happen on the street, then perhaps there is no racial animus. But if you look through data compiled by the Tampa Bay Tribune on cases in which a SYG defense was used, you do not see much of a difference in setting. Some may think that white-on-black shootings are justified more often because it involves the black person as an intruder while black-on-white shootings happen in different scenarios. This is not the case. Black-on-white shootings also occur in the shooter’s home.

None of this is definitive. The answer to the question being asked across America today—would the verdict have been different if Zimmerman and Martin’s races had been reversed—is unknowable. But the available statistical evidence certainly suggests that Zimmerman walked into the courtroom with an advantage that Trayvon Martin would not have had.


  1. Homicides Are A Lot More Likely To Be “Justified” If The Corpse Is Black | Alas, a Blog  ::  7:30 pm on July 16th, 2013:

    [...] And from Roman’s follow-up post: [...]

  2. Two women, two tales, one trial — MSNBC  ::  7:02 pm on July 17th, 2013:

    [...] if—individuals are pretending to ignore it. John Roman, a scholar at the Urban Institute, looked at self-defense killings. He found that cases in which whites claimed to kill blacks in self-defense were far more [...]

  3. Tom  ::  5:28 am on July 18th, 2013:


    Using blanket statistics like those presented here does not accurately portray the realities in the Zimmerman case. You might notice that the white on black and white on white justified shootings are both about 7% higher in the stand your ground states. There is also not a total number in each area which might show skewed statistics. For example, if they only have 10 black on white incidents compared to 1000 white on black incidents the percentages would faulty. I have been in law enforcement for 27 years and one thing I have learned is you can make a statistic read any way you want.

  4. Viewpoints  ::  8:38 am on July 18th, 2013:

    [...] Roman examined Supplementary Homicide Reports submitted to the Federal Bureau of Investigation between 2005 and 2010. He posted his findings on the Urban Institute’s MetroTrends Blog. [...]

  5. John Roman (@JohnKRoman)  ::  1:26 pm on July 18th, 2013:

    Tom, first thanks for your service, it’s a tremendously challenging vocation. I agree with your point that it is necessary to understand both the number of cases and the percentages to understand whether there are racial disparities. I would note that the data I used in this analysis included every homicide in the FBI data over 6 years, so I did not in anyway pick and choose cases. Thanks again for your comment.

  6. Dan Nguyen  ::  9:36 pm on July 18th, 2013:

    John, could you post the raw number of cases that you’re basing your two graphs on? Currently, your piece has not a single reference to an absolute number, which in any statistical study, is a red flag. For starters, this makes it impossible to verify whether your methodology or even your data collection is accurate.

    At first glance, the two charts you’ve posted raise more doubt than actual conclusions. How is it that a graph that removes police-related shootings has such a massive increase in the rate of justifiable homicides compared to a graph that purports to include police-related shootings?

  7. Simpson’s Paradox: the logic of racial disparities  ::  9:22 am on July 19th, 2013:

    [...] Tuesday, we wrote about our analysis of FBI data that describe the circumstances under which fatal shootings are more likely to be considered justified. We described the stark disparity between the rates of justifiable rulings in cases with a white [...]

  8. ‘Mandela: Long Walk To Freedom’ And America’s Fear Of Black Men | News Times Today  ::  3:27 am on July 20th, 2013:

    [...] John Roman wrote on a Urban Institute’s MetroTrends blog progressing this week “Drawing from Supplemental [...]

  9. ASU center produces largest flexible color organic light emitting display | Clothes  ::  5:16 pm on July 21st, 2013:

    [...] Read more… [...]

  10. Police Science Institute  ::  7:04 am on July 25th, 2013:

    While thinking of criminal justice then it should be in a correct manner because many people are getting affected from these justice.

  11. Race, Homicide and Stand Your Ground  ::  3:17 pm on July 26th, 2013:

    [...] two pieces in particular have captured some attention, one published by PBS Frontline, and one on this blog. Many readers have expressed interest in learning more about the methods I used to arrive at the [...]

  12. Black Male Students Voice Zimmerman Trial  ::  11:01 am on August 1st, 2013:

    [...] “The available statistical evidence certainly suggests that Zimmerman walked into the courtroom with an advantage that Trayvon Martin would not have had,” said John Roman in a post on the Urban Institute’s MetroTrends blog. [...]

  13. ‘Stand Your Ground’ laws: Civil rights and public safety implications of the expanded use of deadly force  ::  11:45 am on October 29th, 2013:

    [...] There are racial disparities throughout the juvenile and criminal justice system in America. African Americans are more likely to be stopped and frisked, to have their motor vehicle searched at traffic stops, and to receive longer prison sentences than are whites. [...]

  14. White versus black justice | OUPblog  ::  2:30 am on November 4th, 2013:

    [...] to be self-defense jumps to a staggering twenty-nine percent in non-stand-your ground states and to almost thirty-six percent in stand-your-ground states. This means that whites are ten times more likely to be able to kill a [...]

  15. Tom DesJardins  ::  11:58 am on February 16th, 2014:

    I think these statistics need to be broken down more. Such as where homicides done during the commission of a crime (e.g. robbery), done in self defense, premeditate homicide, domestic violence, etc. Each would have a huge bearing on the adjudication result and could skew the statics depending on what buckets the races fall in. I feel the articles statistics really don’t project anything meaningful without that. Am I missing something?

  16. Nathaniel  ::  9:18 am on February 19th, 2014:

    Great article.

    Feel free to surf to my website :: Home made Energy, Nathaniel,

  17. Wayster  ::  9:53 am on August 13th, 2014:

    You can’t just look at the numbers you have to view the circumstances at which the homicide occurred. We’re the justified killers being robbed abd their life threatened? Most certainly.

  18. Michael Brown and the Ongoing Struggle for Racial Justice | All-len-All  ::  4:21 pm on September 8th, 2014:

    […] Central Park in 1989 and whose lives were upended by this miscarriage of justice. A recent Urban Institute study reports that when there is a homicide with one shooter and one victim who are strangers, a little […]

  19. furniture chick at here  ::  8:01 pm on September 10th, 2014:

    Thank you for any other magnificent article.
    Where else may anybody get that type of information in such an ideal
    manner of writing? I have a presentation next week, and I’m at
    the look for such info.

  20. Harry  ::  6:54 am on November 15th, 2014:

    You can draw many conclusions from this if that is really what you want to do, but the exact circumstances need to be known – that is to say, that if you use these statistics to show that injustice exists, you have to assume that in all cases the perpetrator was in fact guilty – or in all cases not guilty. that is to say, to say that if you do use these stats to show bias, you are saying that the juries made the wrong decision in some of the cases (ie. they either let too many whites go, or convicted too many blacks)- if, in fact the juries got it right, then how does that demonstrate much of anything? This is not a simple situation – most people agree that racial prejudice exists (many kinds of prejudice have and more that likely will for a long time continue) but not to the degree of the past – but there are still those who have to sensationalize it any chance they get — especially those who want to riot after what they call injustice -