Bad evidence: Ten years after a landmark study blew the whistle on junk science, the fight over forensics rages on

I was aware that a good amount of forensic science was questionable, at best, but this provided a lot of insight into what was most controversial and what was happening within the conferences and internal workings of these forensic science groups.

It's a bit scary that there are some "experts" who are nothing more than hacks pushing for the legitimacy of their field simply because that's what employs them, and that this evidence is being admitted and used in court cases to wrongly convict people.

In particular, I found the following deeply disturbing

“How many times do we hear about the planes that land safely at the airport each day? We only hear about the crashes. And then we hear about the same crashes, over and over again, as if the exception is the rule,”

Ignoring the fact that this is just a terrible analogy, let's focus for a second on why we care so much about the failures of forensic science. We care for the same reason that the famous quote

That it is better 100 guilty Persons should escape than that one innocent Person should suffer, is a Maxim that has been long and generally approved.

is so important to many Americans. This isn't just a tangentially applicable quote - this is a quote about the very justice system for which these forensics are being used in.

The fact that

Nearly half of DNA exonerations to date are in cases tainted by forensic errors.

is a statistic that exists at all, should cause us to reconsider just what should be admissible and what shouldn't. We stopped allowing lie detectors because we found out that they simply don't work so why must we continue to use methodologies that don't work. It's like saying your lucky shirt should be admissible in court because 80% of the time you wear it, your team wins their home games.

This is pretty scary when it's used in criminal cases, but there the state has to prove its case beyond all reasonable doubt.

Now apply this junk science to family law cases, where the state sometimes only has to prove its case to the balance of probabilities - more likely than not.

You used to take cocaine. You've stopped taking cocaine and are engaged in treatment. But that shitty hair-strand analysis doesn't take into account the difference between white people's hair and black people's hair, and it says you've used recently. This could be the thing that sees your children taken off you.