Statistics

How would you feel if you were on trial for murder and the jury's verdict depended on whether or not they could really figure out the odds that the DNA found at the scene indicated you were the most likely killer or not? What if you didn't even understand what the DNA expert meant by 1 in 5 versus 1 in 15?

What if a key piece of evidence against you was a generic, plastic garbage bag and one expert claimed the difference between it and a bag found in your possession was miniscule? Then another expert claimed the difference was massive? And no one in the courtroom, including you, could figure out who was right?

Cops use statistics, such as those collected and published by the FBI, to focus their investigations. For example, statistics indicate that most women who are murdered are victims of domestic violence (in other words, are murdered by someone near and dear to them). So, the first "person of interest" in an investigation of a woman's murder is her boyfriend or husband. This is smart investigation--as long as the FBI statistics are right, as long as they're properly interpreted, and as long as the woman wasn't actually murdered by a stranger. It is also imperative that the cops consider all alternatives before they arrest someone.

Prosecutors aren't supposed to use statistics to decide whom to indict, and they aren't supposed to use statistics as evidence or in their arguments--except when the statistics are necessary to interpret forensic evidence. When prosecutors present DNA evidence to a jury, they use statistics. When prosecutors present hair and fiber evidence, they use statistics. When prosecutors present fingerprints and other sorts of evidence, they have to skirt around statistics and talk about "points" that seem to "match" "to a reasonable degree of scientific certainty" (which is a completely unscientific way of stating things, BTW).

The problem is that jurors don't understand statistics, yet they are asked to view statistics as proof of guilt.

For that matter, lawyers and judges don't seem to understand statistics, either. And many expert witnesses do a poor job of explaining statistics to the jury.

One of my pet peeves is the poor quality of math education in this country. I'm not surprised no one understands statistics. I certainly managed to graduate from college without understanding them. I doubt that most people know the difference between the mean and the median. Then there's the whole problem of not knowing how "to do the math" when the math involves percentages.

Two recent trials involved an abuse of statistics. In one instance, the abuse seems to be countenanced by case law--but that doesn't make it right. That trial was the trial of Joshua Rosa. The other instance, it seems to me, occurred because the judge did not understand the math. That trial was the Melanie McGuire trial.

To be continued . . .