The ongoing mystery surrounding the tragic story of Kendrick Johnson led me to wonder whether there is a geographic trend in the ratio of accidental deaths to homicides. Obviously, some states will have higher homicide rates, especially those with large cities. But how does the injury-related accidental-death rate vary? And is the ratio of accidents to homicides controlled solely by population density?

This graphic shows that the ratio is not strongly controlled by population; while less populous states tend to have higher accident to homicide ratios, when I plot population against the ratio on a state-by-state basis, the Pearson coefficient is only 0.15. This suggests there are other factors at play, which is clear just from looking at the accident rate distribution in the graph. What those factors are is beyond the scope of the graphic, but it would be interesting to find a way to determine whether some cities are significantly more likely to rule a death an accident rather than homicide. To do that, we’d need to compare rulings against actual cause of death, and those data don’t exist. Regardless, the graphic shows some interesting trends. The one relevant to the Kendrick Johnson story is that Georgia, at 3.8 nontransport accidental deaths per homicide, is significantly below the national average of 5.3.

Data source: http://www.cdc.gov/nchs/data/nvsr/nvsr61/nvsr61_04.pdf

Posted on 16 October, 2013

5 notes

  1. natalierushmanuspdf reblogged this from vizual-statistix
  2. jarrodlombardo reblogged this from vizual-statistix and added:
    "we’d need to compare rulings against actual cause of death, and those data don’t exist." In locations with a Medical...
  3. vizual-statistix posted this