The Washington Post collected data on more than 52,000 homicides over the past decade in 50 of the largest American Cities 1. The cities were selected based on their size and their violent crime rate as reported to the FBI.
They analzyed the data by geographic area and found disparities between arrest rates.2 Baltimore and Chicago had large areas where homicides mostly went unsolved. Atlanta and Richmond, on the other hand, had few areas where murders where unsolved.
The data was made available. 3 I looked at the data by race. Blacks are victims of homicide much more often than any other race, and their murders go unsolved at a much higher rate.
Having just finished a course on Inferential Statistics 4, I performed a chi-square test on arrest disposition and victim race and found a p-value of 1.311608e-251, essentially a zero chance that race and disposition are unrelated.
| statistic | chisq_df | p_value |
|---|---|---|
| 1200.189 | 10 | 0 |