Impressions from Chicago
I survived my first day at the Academy of Forensic Sciences Convention. It was nothing like anything I have ever been to. Having been to a number of national seminars, I've seen some pretty big crowds - but nothing like this. I have no idea how many people there are, but I'm guessing well over a 1,000.
The seminars are set up a lot like TCDLA's annual forensic seminar, with a number of tracks going on at the same time. Basically there are 12 different seminars going on; which means you have to spend a lot of time mapping out what you want to go to. A big difference from most legal seminars I have been to is the length of the presentations - most are 20 minutes. I supposed most people would say a lawyer can't introduce themselves in 20 minutes - and there is some truth to that.
Much of today was spent discussing the problems with forensic evidence; not surprising given the attacks that have flowed from the NAS report. As you can imagine, some are none to happy with some of the tactics being used. However, everyone I listened to today acknowledged that there are problems that need to be identified and addressed.
One of the first sessions identified the "monopoly" in forensic science. The idea is that the vast majority of forensic tests are done by crime labs; most of the testing is never challenged, wihich results in a monopoly in interpretation. Obviously if something is never challenged, the likelihood of uncovering mistakes is small.
Most of he presenters recognized the concept of cognitive bias. One referred to it as "observer " bias, or "rational bias", noting that it always exists. As he said, everyone has a point of view - you can't avoid it. The problem arises when tests are ambiguous - you have to choose how to resolve that, which cannot help but be influenced by your point of view. When you work every day with police and law enforcement no one should be surprised if you have a favorable view of them - and want to help. (this is my own theory - not anything anyone said today by the way)
Another big topic of discussion was error rates, with most people pointing out how non-specific it is; in other words, what does it mean. A more accurate analysis would be to look at the number of false positives (how accurate the tests are), and false negatives - how precise or specific they are. Another important concept is resolution - which is how precise a test is. The example is ABO vs DNA testing - where one includes someone and the the latter excludes them. The ABO test was correct in including the person, even though the more precise test excluded them. An interesting paradox is that the more precise a test, the more chances there are for error - think of a ruler. The more measurements there are the greater the chance you will read it wrong.
A couple of professors in statistics/mathematics also presented. They presented the concept of a likelihood ratio - which shows how likely the test results are correct. I'll talk more about in another post.
There was also some good information on pathology and trace evidence, as well as DNA. A recurrent them was that lawyers need to get all the information possible. There is much more interpretation involved in all areas of forensic science than most people realize. An interesting discussion addressed the reporting of DNA results involving a mixture - there were several opinions regarding how it could be reported. All were correct, and all involved choices. Unless you had all the information you wouldn't know what choices were made, and whether it was a problem.
Tomorrow's another day. I'm sure it will be just as useful.
