FBI admits flaws in hair analysis over decades (2015) | Hacker News Reader

FBI admits flaws in hair analysis over decades (2015)

washingtonpost.com

80 pointsrbritton6 comments

Previous discussion last year here[0] but I had completely missed that, so thought it would be of interest again given the current events.

[0]: https://news.ycombinator.com/item?id=9401453

One of the biggest issues with thinking innocence until proven guilty beyond a reasonable doubt only applies to courts is that courts are not divorced from public opinion. The whole 'tough on crime' stance happens because in public opinion, a guy is guilty with far less evidence than the court needs. If the court cannot also find them guilty, the public sees this as a problem. Nothing direct happens, but as this occurs time and time again people begin to vote for those who don't allow this 'miscarriage of justice' to continue to occur. So over time, the public selects for people who act in ways that do not support innocent until proven guilty beyond a reasonable doubt. DAs willing to ignore evidence. Judges willing to present cops as being more trustworthy a source (such as allowing them to show up in uniform). And even crime labs who are a little willing to help convict the guy who is 'obviously' guilty.

In short, because public opinion eventually impacts the courts through some path, if the people don't believe in innocent until proven guilty beyond a reasonable doubt in the case of public opinion, then the courts will eventually lose that focus as well.

I am not a lawyer and I don't know how federal court works so I'm genuinely curious:

FBI forensic "scientists" are essentially always lying on the stand -- not out of malice or intentionally, but because their techniques do not have nearly enough accuracy to honestly present the results that they present. There is plenty of evidence for this, this article and others. Anything except DNA is essentially a fortune teller reading a palm.

So why does every single defense attorney in the nation (including public defenders) not spend hours on this particular fact any time an FBI forensic scientist testifies in a case? And essentially make having forensic evidence from FBI crime labs a huge liability and unusable in court?

Do they try and people just believe the FBI anyway?

>Anything except DNA is essentially a fortune teller reading a palm.

DNA evidence has been somewhat problematic as well, though mostly due to improper/non existing lab standards:

https://www.utexas.edu/courses/bio301d/Topics/DNA/text.html

In the earlier days of DNA analysis, they routinely overstated the certainty of matches. They were testing too few loci for the kind of results they were claiming.

Matches that were called something like one in 8 billion were sometimes more like one in 2 million. In a city the size of New York, that's far less certain of match than they stated.

It is also bad statistics.

The expert points out that it is 1 in a million. The jury will think this means that there is a 99.9999% chance this guy is guilty.

But how it works (assuming the DNA is the only evidence to connect this guy to the crime) is that in a city of 100 million, there are 100 possible people who would match, so the actual chance this one guy who was found by DNA match is actually the guilty one is only 1%.

People are willing to convict if given 99.9999% certainty. But who would be willing to convict with only 1% certainty?

Reply on news.ycombinator.com