I read a news story about the use of statistics in the criminal justice system. The case involved the murder of Erik Tornblom by a man named Marcus Robinson. Erik was 17-year old high school student who worked at a restaurant in North Carolina. After closing up the restaurant at midnight, Erik was about to get in his car to drive home.
Robinson and an accomplice kidnapped Erik to rob him. Erik had no money and begged for his life, but Robinson and his accomplice shot Erik in the face with a shotgun, killing him in a premeditated and gruesome way. A few days before the murder, Robinson told a family friend that, “He was going to burn himself a whitey.” Robinson was identified, and arrested. Robinson confessed to the murder. Robinson was convicted and sentenced to death by a jury that had both Black and white members. That should have been the end of the story.
Left: Erik Tornblom was a 17-year old high school student who was killed by a shotgun blast to his face. Center: The convicted murderer, Marcus Robinson, celebrates his release from death row, due to a ruling by the North Carolina Supreme Court. Right: The North Carolina Supreme Court might benefit from an explanation of basic statistics.
OK, so what does this have to do with statistics? The State of North Carolina had passed a law called the Racial Justice Act. It states loosely, that if there are any statistical differences in the rates at which Blacks are convicted in any county or in the State as a whole, a convicted Black murderer can request a more lenient sentence.
The North Carolina State Supreme Court ruled that because Blacks are convicted of murder more often than white people in North Carolina, and because there are more employed white people on juries than employed Blacks, Robinson should be granted release from death row, and be re-sentenced to life with an outside possibility of release/parole.
This is a poor use of statistics for several reasons. The North Carolina law doesn’t apply to any particular case; it’s a statistics law. Put another way, the law doesn’t take the details of the Erik Tornblom case into account at all. It’s also an example of correlation-is-not-causation. There is overwhelming evidence that Blacks are convicted of murder more often than white people because they commit more murders than white people. This fact shouldn’t be relevant to Erik’s murder.
The overall moral of the story is that you shouldn’t apply statistics to an individual person or case. Statistical data deals with numbers. Not people.