'Great Pause' Among Prosecutors As DNA Proves Fallible
npr.org
npr.org
I often hear people say this, but do you actually have the ~20-200+k to actually sue in this case? Because, the vast majority of people out there really don't have the option to just sue. And, I think that's part of why are justice system stays so broken. Many people only find out how corrupt it is after there in it's cross-hairs.
Or, at least getting every piece of "non-science" evidence thrown out?
Having a public defender represent you when they have hundreds of other cases concurrently is useless.
This could turn into a good thing (for the defendant). Simply argue that the state provided you with incompetent representation and therefore any continued proceedings by the court are unconstitutional.
In fact, I don't know why that is not currently the case. Surely there must be some blind testing, and processes to stop, for example, a technician testing the blood samples on their wife's murder.
So I would be surprised if most of what you say is not already true, but also surprised if it werent
- The prosecution, police and crime lab are solely focused on conviction, and
- the judge is like a referee, focused on procedure
It's like "trial by combat" in Game of Thrones. Alternatives exist. Germany has an inquisitorial system where judges actively investigate the facts of the case.
https://en.wikipedia.org/wiki/Judiciary_of_Germany
A little bonus crazy: in the US prosecutors have absolute immunity for anything they do during the prosecution.
- The way this is practiced is may however differ depending on the particular prosecutor.
This is an individual problem rather than there being an inherently adversarial system.
[1] https://www.americanbar.org/publications/criminal_justice_se...
It's both. My guess is that the vast majority of prosecutors do think about justice as well as about winning. But consider the incentives:
+ All lawyers tend to want a high win rate, because wins serve as a marker of professional competence. Wins help to bring in more work --- no one wants to have to sit around waiting for the phone to ring --- as well as more-prestigious and better-paying work.
+ Prosecutors in particular want the Ws to enhance (i) their chances of promotion within the government system (and, in the case of elected district attorneys, their chances of re-election) as well as (ii) their ability to move to lucrative private-sector jobs as defense attorneys, or even a TV gig as an anchor or reality-show host.
For example, the Houston Chronicle recently reported that the assistant U.S. attorney who successfully prosecuted financial fraudster R. Allen Stanford just moved to the giant Jones Day law firm to do white-collar criminal defense work. You can be sure he'll earn far more money in private practice than he ever did in government service. [1]
As another example, former Houston state prosecutor Kelly Siegler is one of the hosts of a cold-case-files series on the TNT cable-TV network; it's a safe bet that she'll make more money doing that than in her old job. [2]
[1] http://www.houstonchronicle.com/business/article/Jones-Day-s...
[2] http://www.houstonchronicle.com/entertainment/tv/article/For...
Some city's do it this way, and it makes a big difference. Otherwise forensics budgets are always competing with "more policemen". Guess what, policemen win almost every time.
Although the headline is highly editorialized, it's clear that DNA evidence is just as fallible to bad science as any other form of evidence and that we must treat all forms of evidence with respect and skepticism if we are to find the truth.
1 in 40 false matches rather than one in a million, throughout thousands of cases? Outrageous.
However, since DNA analysis techniques are mutable, it is possible for DNA analysis to be useful, as long as it is used correctly.
I believe there is a difference.
Illumina is pursuing mitochondrial DNA (mtDNA) sequencing. An advantage here is that cells contain many copies of mtDNA sequences (as opposed to just one copy of each gDNA haplotype), and mtDNA contains hyper-variable regions which confer strong individual specificity. This is potentially advantageous in crime scene samples, where the DNA could be damaged through degradation processes like sun exposure. http://www.illumina.com/areas-of-interest/forensic-genomics/...
Ion Torrent/ThermoFisher are going after the same STR targets, but using their Torrent and Proton NGS platforms (rather than PCR). Unlike regular PCR methods, this can provide things like allele frequency estimates, and can call more than one base into variable regions (which provides more information, and can potentially be used to infer things like height, ethnicity, hair or eye color). https://www.thermofisher.com/us/en/home/industrial/human-ide...
Carlos Bustamante (a Stanford Professor, and world expert in ancient genomics / diverse population genomics: https://med.stanford.edu/profiles/carlos-bustamante) has founded IdentifyGenomics, which is a startup focused on new methods for forensic DNA sequencing (disclosure: I know Carlos, but I'm not involved in his startup).
Definitely an important problem, and will be interesting to who succeeds in converting forensic investigators to use NGS at scale.
One specific answer is that the FBI is gearing up to regulate new devices in this area (https://www.fbi.gov/about-us/lab/biometric-analysis/codis/st...). For example, in this publication (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3757157/), they specifically analyze an IonTorrent PGM for use in forensics applications, which we'll probably see them do for various other platforms that come on to the market.
Separately, and outside the forensics realm, there's a trend towards increased regulation of DNA sequencing. For example, NIST has developed/is developing methods to evaluate sequencing platforms: http://www.nist.gov/mml/bbd/dna-022514.cfm. This is relevant to other sequencing applications too (e.g. personalized medicine, somatic tumor profiling, etc). The FDA are also involved here, but more focused on medical applications.
So, I think collectively through both the increase in the forensics community regulating forensics NGS applications, and more broadly the biomedical science/technology community regulating general NGS platforms, we'll see good technology validation standards (at least in the U.S.). But the significantly higher complexity of these systems does introduce more opportunity for error, so it's entirely possible we'll see similar biases in NGS based forensics.
It's scary to think prosecutors have been pitching DNA as infallible, damning evidence all along. It means getting a defense attorney off their game might very well lead to an incorrect conviction (think Serial, though obviously the circumstances there are more tied to the use of the cell phone records). I don't think that technology is going to bring us easier convictions for a while yet (e.g. statistical analysis of evidence leading to inference ala Watson might be interesting). Somehow this isn't reflected in how juries have chosen, though, so I think public education about it (even via entertainment) might be the most effective way to change this.
It takes some technology that most people don't have (like a centrifuge) but the actual process is simple and can be achieved by anyone with familiarity with the equipment itself.
This could be a case for using shotgun / whole genome sequencing - we'd expect to see a more even distribution of genome coverage than one would get by leaving targeted oligos behind. But this isn't likely to happen anytime soon; the costs are far too high (10-100X greater than targeted sequencing).
In the meantime though, one highly feasible avenue for spotting synthetic DNA oligo fragments is the presence and position of nucleosomes (https://en.wikipedia.org/wiki/Nucleosome), which are DNA-associated protein complexes that occur in DNA from organisms, but not in synthetic DNA oligos. These have been used successfully to trace tissue of origin in cell free DNA in humans, and have specific signals related to chromatin/genome topological state. The only way I can think of to fake this signal would be to have a cell culture from the person you're trying to imitate. Granted, this too is not completely unreasonable - there are now very robust commercially available protocols for deriving iPSCs from small dermal fibroblast samples.
Juries are largely representative of the general population. As such, most have been conditioned by television crime dramas such as Law & Order, CSI and NCIS to view DNA evidence as infallible techno-wizardry that always catches the bad guys.
Ever see a mainstream crime drama end with the arrest and conviction of someone totally innocent, the credits roll, and that's it? Neither have I.
Instead, we have the stereotype of law enforcement being completely incorruptible, with the evidence technicians themselves cast as well-meaning nerds who are equally infallible, because gosh, they're just super geeky scientists who are super smart.
"When you go into court you are putting your fate into the hands of twelve people who weren't smart enough to get out of jury duty." - Norm Crosby
Sadly that quote is even more relevant today.
Justice https://en.wikipedia.org/wiki/Justice_(2006_TV_series) is a legal drama that sometimes the innocent are convicted or the criminal walk free. It is a fantastic show all around in my opinion, but sadly it got cancelled after one season due to low ratings. Perhaps that fact validate your position even more, in that people don't like to face the fact that the judicial system is fallible.
Occasionally large-scale corruption comes to light such as when a Boston crime lab falsified drug tests on a massive scale [2] or New York state police were found to have fabricated fingerprint evidence for nearly a decade [3]. The "forensics" field has been embroiled in a steady stream of scandals [4]. Far from being isolated incidents, this is the norm within law enforcement.
[1] http://www.tandfonline.com/doi/abs/10.1080/0731129X.2013.817...
[2] http://www.policeone.com/csi-forensics/articles/5956534-Mass...
[3] https://en.wikipedia.org/wiki/New_York_State_Police_Troop_C_...
[4] http://www.abajournal.com/magazine/article/crime_labs_under_... http://www.theatlantic.com/politics/archive/2015/04/csi-is-a... https://www.nacdl.org/criminaldefense.aspx?id=28286 http://listverse.com/2015/02/06/10-heinous-cases-of-miscondu...
Seriously: what?
Funding crime labs through court-assessed fees creates another channel for bias to enter crime lab analyses. In jurisdictions with this practice the crime lab receives a sum of money for each conviction of a given type. Ray Wickenheiser says, “Collection of court costs is the only stable source of funding for the Acadiana Crime Lab. $10 is received for each guilty plea or verdict from each speeding ticket, and $50 from each DWI (Driving While Impaired) and drug offense.”117 In Broward County, Florida, “Monies deposited in the Trust Fund are principally court costs assessed upon conviction of driving or boating under the influence ($50) or selling, manufacturing, delivery, or possession of a controlled substance ($100).”118
Several state statutory schemes require defendants to pay crime laboratory fees upon conviction. North Carolina General Statutes require, “[f]or the services of” the state or local crime lab, that judges in criminal cases assess a $600 fee to be charged “upon conviction” and remitted to the law enforcement agency containing the lab whenever that lab “performed DNA analysis of the crime, tests of bodily fluids of the defendant for the presence of alcohol or controlled substances, or analysis of any controlled substance possessed by the defendant or the defendant's agent.”119 Illinois crime labs receive fees upon convictions for sex offenses, controlled substance offenses, and those involving driving under the influence.120 Mississippi statues require crime laboratory fees for various conviction types, including arson, aiding suicide, and driving while intoxicated.121 Similar provisions exist in Alabama, New Mexico, Kentucky, New Jersey, Virginia, and, until recently, Michigan.122
Other states have broadened the scope even further. Washington statutes require a $100 crime lab fee for any conviction that involves lab analysis.123 Kansas statutes require offenders “to pay a separate court cost of $400 for every individual offense if forensic science or laboratory services or forensic computer examination services are provided in connection with the investigation.”124 In addition to those already listed, the following states also require crime lab fees in connection with various conviction types: Arizona, California, Missouri, Tennessee, and Wisconsin.125
Glen Whitman and Roger Koppl point out that “the very choice to submit a suspect's sample to the lab makes the lab more inclined (than it would be otherwise) to announce a match, indicating that the suspect is guilty.”126 The forensic scientist must evaluate ambiguous evidence, but give, generally, a binary judgment that the evidence does or does not match. Whitman and Koppl explain why the probabilities given in DNA testimony are not usually an exception to this binary nature of forensic-science testimony.127 In this situation, even the most rational scientist must choose what to say. The choice will usually be influenced by scientific analysis done in the crime lab. But if the evidence is ambiguous, as it often is, then two other factors matter even for perfectly rational forensic scientists. The scientist is more likely to inculpate the defendant (1) the higher the forensic scientist's “prior” probability of guilt, which is the probability before the forensic evidence is examined, and (2) the weaker the scientist's desire is to avoid convicting the innocent relative to her desire to convict the guilty.
So, the claim does appear to be true, at least in some states.
[1] http://www.tandfonline.com/doi/full/10.1080/0731129X.2013.81...
> "When they retested it, the likelihood that it could be someone else was, I think, one in 30-something, one in 40. So it was a significant probability that it could be someone else," Torres says.
No. That is the probability that the test would have claimed a match with a particular person, conditioning on it not being the right person. In order to calculate the probability that it was someone else, you need to know the probability of it having been the right person prior to the test being performed. If you take a sample and test it against everyone in a big enough database, with a 1-in-30 false positive rate, then you will definitely end up accusing someone falsely.
'...threatens to undermine the deep faith people have placed in the technology.
'"And it's not faith they should not have had to begin with," says Keith Inman, who teaches forensic science at California State University, East Bay.'
And that's clearly wrong. Someone without sufficient statistical ability and without any incentive to look for problems dreamed up a protocol and it was accepted (with "deep faith" no less) without justification.
And the academics who should be the most disinterested seekers after truth isn't willing to call this mistaken, but we're supposed to pretend no one could have seen this coming.
[1] I am pointing this out because HN readers are presumably more intelligent than "most people" on average.
[2] You can find a lot of comments like that in this post: https://news.ycombinator.com/item?id=9629797
Faith in technology or people who work for and are paid by the police and who know that DNA matches are good for repeat business? It's a complete scandal that DNA testing does not follow a line-up approach: here are 100 samples, one may be the suspect, tell us which one if any matches the evidence. So called partial/mixed matching will become the hair analysis scandal of this decade.
A systematic miscarriage of justice that has ruined many thousands of lives. And the NPR identifies the "victim" here as the lab and the prosecutors.
I'm not quite sure if the situation around the world is much different, but I would believe that we're a little more sane in Europe when it comes to admitting evidence based more on witchcraft than solid engineering.
The analog here is there are M people in the database, and N crimes, so the chance of a single failure (mistaken match) is kind of like N*M.
But I'm not sure that's the correct metric; I think you really care about the amortized cost of a mistaken match across all N crimes. So you don't see the "paradoxical" quadratic increase.
I admit I'm being a bit fussy -- just testing the extent of the analogy.
You are of course correct that the database size is growing, and the number of samples that can be collected, even from one crime, also can be made to grow.
It seems to me what we should really care about is the number of people falsely convicted, which is false-positive-rate-per-defendant * number-of-defendants. Since the size of the database (and therefore the false positive rate) is related to the number of defendants, the number of falsely convicted people increases quadratically with the number of people subjected to DNA testing.
As an aside, even if "one in a million" were true, we're talking about 300+ matches in the US and 7000+ in the world.
IMHO, then the correct question to ask is: "given there are 300 people in the US whose DNA matches the one of the murderer, how likely is it that the defendent is the murderer?"
So you're comparing a suspect's DNA with DNA from a crime scene. And you find the same set of marker sequences. But maybe the crime-scene DNA is a mixture. However, with a mixture you'd expect to see additional marker sequences in the crime-scene DNA.
But it depends. Many of the marker sequences are racial. So maybe it's more likely for two people of a given race to look like one person. And maybe there was racial bias in marker selection. It's a mess.
Bottom line, with current technology, people should not be convicted based on the occurrence of marker sequences. Nothing less than complete DNA sequences should be admissible.
Now, npr, they always, always, ALWAYS take the side of the underdog, no matter the less favorable actual odds. They put more faith in soft science results (child psych says this is good for kids) than results from scientific studies (study showed none to negative impact).
Example, people in Seattle placing signs saying No Californians, it gets a chuckle, could you imagine a similar sign saying "No [insert poor country demonym]. NPR would be outraged, with reason, but you don't see the outrage when the same bias is projected against the relatively better off.
I would not extrapolate the error in the headline across the entirety of NPR, though. The reporting in this article was factual and (AFAICT) unbiased. Although I believe that "soft sciences" as opposed to "hard sciences" exist, and that NPR can be biased (they have been in the past), I do not believe this discredits NPR entirely. The mistake they are reporting allegedly affects thousands of cases. That statement is unlikely to be a result of bias.
It is very hard to prevent bias at the publishing outlet level (and very easy to introduce it). I recommend that people get their facts from a number of sources and that they listen to every opinion they can find.
I think one other thing we can take from this is that DNA alone is insufficient to establish guilt, we need corroborating evidence. Just because someone's DNA was probably on a knob does not imply that person was the burglar. Maybe they were a repair person, but they could have been a burglar, if they returned afterwards and broke in, for example. Of course lawyers will point that out, but sometimes people think DNA = guilt.
The issue with npr is they like to have the image of being unbiased. If you can suffer through their pledge drives, you hear their listeners fawn and claim them to be the purveyors of truth and unbias and they, the journalists and interns, truly seem to believe they are.
Lets say an underdog commits a crime, something if done by a top dog, they would take it at face value) They did the crime, jail them. If it's an underdog, well, maybe there were deficiencies in society which contributed to their criminal behavior and thus perhaps that should be taken into account when they are tried.
It's worth pointing out that journalists almost never get to write the headlines for their articles. Headlines often drive them crazy.
Oh give me a break. NPR is a standard mainstream news organization, with a higher journalistic standard than most. They don’t by any means “always always ALWAYS” do anything, except try their best to discover topical news stories and report them fairly. Like other mainstream news organizations, they aim for political neutrality, tend to analyze stories in line with the “consensus”, and shy away from saying anything which deeply challenges the dominant power structure.
You might argue that standards of evidence in science-related journalism are weak across the board, but let’s not pretend that there’s some problem with NPR in particular that wouldn’t equally apply to, e.g., the New York Times, the Associated Press, the BBC, Le Monde, or the Economist.
If you set your bar this low, did you have to dig a trench first? What did you do when you hit the water table? Did ground heat ever become a problem as you edged closer to the Mantle?
Besides being a contradictory statement, I'd like to see an unbiased, reputable source regarding the second half of your comment.
"X is a Y Z, with a greater W than most" I think would generally be interpreted as "X is a Y Z with a greater W than most Y Zs"
So, X would have a more than 50th percentile ranking in Y Zs, when ranking by W.
I don't see why it would be impossible for a standard mainstream news organization to have a greater than 50 percentile ranking when ordering standard mainstream news organizations be journalistic standards.
Clever bit going all-variable when there is no frame of reference, starting point, or baseline.
On its face, if it is a standard, mainstream organization, it has srandard, mainstream standards. However, there may be some source that indicates otherwise (the call in the second half of my statement).
Nope. False. Grotesque, outrageous, brain-dead incompetence. Instead, even with the "certainty" of one million to one, might find a "match" in the next person or not in 10 million people.
Also, where'd they get this certainty stuff? I have an excellent background, thank you, in pure and applied probability, and we don't use certainty like that or, really, hardly at all.
Police work with DNA and associated probability calculations? This is from a Saturday morning TV rerun of some old movie of The Three Stooges, right?
With me on a jury, as soon as the prosecutor presents evidence from a police lab and/or collected by the police, especially DNA evidence, I'll be tempted to stand in the jury box and call for acquittal via acclamation, dismissal of the prosecutor and police for incompetence, and their prosecution for fraud!
As soon as a prosecutor presents DNA evidence, I know he's talking total nonsense and, then, tough to get a conviction.
Or, commonly a prosecutor will want to show that the defendant has a background of lying, etc. Well, as soon as a prosecutor presents DNA evidence, I have to strongly suspect that he is lying or at least incompetent. Can't convict; have to acquit.
(1) Police departments? (2) The biochemistry and probability theory of work with DNA? Two things that should never be combined!
Police? Working with advanced, delicate topics in science and math? What a joke -- what The Three Stooges could make of that. Now definitely material for SNL!