The National Academy of Sciences has released a report stating that the Forensics community needs an overhaul. There isn't enough regulation or research and standards of practice aren't uniform across the community. Unfortunately, the majority of this is true. I once read that when a handful of Forensic Odontologists were asked to match teeth impressions to sets of teeth, they only got it right about 60% of the time. I don't know the details of this test, but regular proficiency testing would weed out those who couldn't get the matches. And what constitutes a match? How much similarity does there have to be? And does a "match" match a person with a piece of evidence? For example, does a carpet fiber at a crime scene similar to a carpet fiber in an individual's home place a member of that particular household at the crime scene? Essentially, these test results needs to be qualified.
But DNA testing is different - maybe because it had the good fortune to come into being after most other forensic disciplines. DNA testing is regulated by several sets of industry standards. Laboratories are requiried to be audited and accredited. DNA analysts must be trained in a specific way and proficiency tested semi-annually. And the cherry on the cake is that any DNA match must be qualified when reported. This is done using a statistics database that catalogues the frequencies of DNA genotypes in the general population. Statistical calculations are performed to evaluate the chance of a random person possessing a given DNA profile. In other words, the jury is given an idea of how likely it is that this DNA match is coincidental. It weighs how damning the evidence is.
In summary, DNA testing stands out in the NAS report as the one branch of Forensic testing that's doing it right. With the help of organizations like NIST and guys like John Butler, we will continue to keep doing it right.