John Roman emails the Dish:
Many thanks for featuring my data. I noted the critiques you highlighted and wondered if I might respond. A reader was concerned that the small number of cases that look like the Zimmerman
case nullify the analysis. To clarify, the 23 cases are cases where there is a single victim, a single shooter, they are strangers, neither is law enforcement, a handgun was used and the shooter was older than the victim. It’s not a sample of those kinds of cases – it’s every single one in the five years of FBI data. To generate the data in the chart, I ran a regression model, which includes about 5,000 other cases that did not share all of those attributes, so the overall sample is about 5,000, not 23, which is great for a social science dataset. There is no data about income in the dataset and unfortunately nothing about the setting where the shooting occurred (residence, business, street), but I controlled for everything else available.
The bottom-line is, I would have preferred to have conducted the data analysis your readers were looking for, but those data simply are not available anywhere. And I agree with them that my analysis is not sufficient to make any causal statements.
However, I do note that in a criminal justice system rife with disparities – blacks are disproportionately more likely to be stopped and frisked, to have the cars searched at a traffic stop, to be convicted and to receive longer sentences – this disparity dwarfs them all. So, it’s certainly worth discussing.
Robert VerBruggen joins the discussion:
I saw your post about the Stand Your Ground chart from the Urban Institute, and I just wanted to note a few other points I made at RealClearPolicy. Specifically, the overall racial difference (i.e. white-on-black homicides being more likely to be ruled justifiable in all states, SYG or not) might be explained by two factors – one, according to FBI data, black/white violent crimes are more likely to involve black offenders and white victims, so whites are presented with more opportunities to commit justifiable homicides against blacks than vice versa; and two, whites are more likely to own guns, so they’re more likely to have the means to commit justifiable homicide when they have the opportunity.



of justifiable homicides between the races. It was troubling enough that I decided to track down the source of the study. In doing so, I found out an even more amazing statistic. The study looked at all 73,000 homicides between 2005-2009. It then separated all of the homicides where one stranger killed another stranger, similar to what occurred in the Trayvon case. Finally, the study separated those instances by race. Of the 73,000 homicides in that time period, only 23 were one white person killing one black person. 23! From the media coverage this weekend, I thought the number was probably in the thousands. The small sample size makes the significant portion of the graphic you posted basically worthless.




