|
It can eliminate with a high degree of certainty, but cannot identify with anywhere near the same certainty. The process used does not sequence the entire genome, but instead looks at key areas. Hence, while elimination is very easy to do (one key area does not match), if all match, what do you know?
Well you know that the "odds" are about 1 in 6.25 million that you misidentify. However, those odds are relative to the number of people in the world, not at a potential crime scene. Hence, the odds are misleading, and insufficient.
The real problem, though, is not in using the technique. It is the air of infallibility associated with a "scientific" process that easily sways opinion, both inside the jury room and outside in public and official opinion. People want to believe in a conviction, and DNA is "scientifically", or at least technically, mysterious and seemingly foolproof.
Randomly taking DNA and then "matching" it to then accuse someone of a crime quickly turns those 1 in 6.25 million to the real odds of of near 1 in 1 that you will accuse a large number of people inaccurately.
As always, the way to approach understanding odds is not in restating them as likelihoods in general, but in specific terms. E.g., the silly argument that evolution could not have happened because "what are the odds" that humans just like us could have evolved. After all, we are quite unique, complicated organisms, not to mention the rest of life on Earth. But that is the wrong question. Eveolution is not posulating that natural selection leads to THIS EXACT RESULT on Earth, but that it will inevitably lead to a state that is in agreement and explains where we are now.
So, back to DNA evidence, the question is not, in this case, whether a handful of people "match" the DNA signature measured in evidence, it is instead of the 6-7 billion people on Earth, or in this case maybe only the 10-20 million people who have had their DNA "recorded" in the U.S. to "match" against, whether ANY of them "match." Suddenly the 1 in 6.25 million odds looks like the odds on favorite. OF course some of those people will "match." In fact, as we collect more and more samples as electronically archive the "fingerprint" and then "match" them up, there will hundreds and then thousands of matches, and thus more opportunity to harass the people you have arrested.
Just perfect.
|