Monday, March 12, 2018

Pop quiz on forensics

A recent research paper from Jonathan Koehler, a Northwestern University law professor, titled "How trial judges should think about forensic science evidence," opened with a short quiz. Grits readers should be well educated on these topics, but let's see how you do:

"Here is a forensic science test for you. Please answer each of the three questions below True or False.

"1. Scientific tests conducted over the past 100 years have repeatedly demonstrated that everyone has a unique set of fingerprints.

"2. Recent scientific studies show that the chance that DNA samples from two different people will be identified as a “match” by a competent, well-trained DNA examiner is less than one in a million.

"3. Data from scientific tests conducted over the past few decades provide a reliable basis from which to estimate the accuracy of most forensic methods that have been admitted in U.S. courts.

"The answer to all three of these questions is False. How did you do?"


Zoe said...

Even more problematic, these deficits been known for the past 30 years, yet there has been very little movement. Mostly because of very little accountability for using shoddy science.

Anonymous said...

One of the first things that a trial judge should consider when thinking about scientific evidence is whether another lawyer is the best source of information about science. Question 1 is case in point. Uniqueness is not a characteristic that can possibly be demonstrated as a scientific certainty. It is the lawyer’s fantasy of what science should be able to say, but it is not what science really says.

Gritsforbreakfast said...

Funny, 9:11, science can say DNA is unique to a scientific certainty. Perhaps the reason it can't be done with fingerprints isn't the limits of science but the non-scientific nature of fingerprint evidence?

Steven Michael Seys said...

I read an article a couple years ago that told how British scientists discovered two people whose finger prints match with each other under every known forensic standard, yet they were not related nor even from the same continent. But the forensics isn't so much the problem as the presentation of that as testimony in court. In my trial, the prosecutor got the ME to tell the jury that all the victim's knife wounds were inflicted by the same 8-inch cook's knife, but the autopsy report describes one wound as 5/8" entry and 3" depth. At 3" the width of the cook's knife is greater than an inch and therefore it physically could not have possibly made that wound. This is the kind of forensic testimony routinely delivered in courts. If the jury were competent to read and understand forensic reports for themselves and such reports were then actually compared to the testimony at trial these discrepancies will be exposed and the state's witnesses no longer given greater trust than the non-professional witnesses of the defense.
I am unable to cite the article at this time because a couple of years ago I lived in a five foot by nine foot apartment with a roommate in a condominium run by the state of Texas. I'll try to find it online whenever I can get WiFi.

Anonymous said...


9:11 here.

Don't get me going about lawyers and DNA. But that said, here is my nutshell rant.

"Unique" has a particular meaning. A characteristic that is unique is a true one-off - never to be seen again, neither now, before, nor in the future.

A characteristic that is "unique to a scientific certainty" is not unique - it is rare.

At the very, very best, what can be said about a forensic DNA profile (the best quality DNA profile from a single individual that is not degraded and not low-level) is that it is rare in a particular population group of interest. The numerical degree of rarity can be estimated (yes, estimated). And that numerical estimate might indicate a degree of rarity that would allow a trier of fact to reasonably conclude that a particular sample truly originated from a particular specific person. But when the trier of fact reaches that conclusion, he/she is stepping outside the limited bounds of what the science says - which is that the profile is rare. The trier of fact is making a real-world deductive conclusion using scientific information as the basis for the deduction. But, it is still a deduction, not a fact.


1) It is possible for two unrelated people to share the same forensic DNA profile. But it would be a rare occurrence.

2) It is also possible for two siblings to share the same DNA profile. It would also be a rare occurrence, but way more common than in unrelated people.

3) It is also possible for two first cousins to share the same DNA profile. It would also be a rare occurrence, but less common than in siblings, and more common than in unrelated people.

Of course, that is for the best, most complete forensic DNA profiles. For a sample that gives a partial profile (due to the DNA being degraded or low-level) the rarity estimate would fall off precipitously. The profile might be so common that a trier of fact would be poorly advised to conclude that the sample originated from a particular person.

Everyone, but especially lawyers, would like science to provide certainty to the messy legal decision-making process. But, science does not provide certainty. At best, science provides a quantifiable measure of uncertainty. Which is a very different thing.

Will Yablome said...
This comment has been removed by the author.
Gritsforbreakfast said...

I can grant all that, 5:56 (without fact checking every jot and tittle - first cousins sounds far-fetched), but perhaps you'll grant that the mathematical certainties behind fingerprint evidence are far less exact. DNA is analyzing and measuring an objective thing. Fingerprint evidence relies fundamentally on the subjective interpretation of a trained examiner who works for the cops, and is a technique created by law enforcement to accuse people, not something developed through application of science or the scientific method. That's a different kettle of fish from DNA.

Anonymous said...

Ignorant Judges are only a very small part of the problem.

Of a much greater concern is that greater that 95% of cases don't see the inside of a courtroom, so the science is not vetted for a judge and jury. An official lab report (outside the courtroom) that finds blood/drugs/semen/DNA (and false positives) is strong support to elicit a guilty plea, even for defendants who are factually not guilty (i.e. coerced).

Of those lab reports that do actually make it to a courtroom, the focus is only on the scientific report of that specific case on that specific evidence at that specific time, and not the thousands of other cases that are affected by long-standing lab systemic problems. (e.g. As a prosecutor might declare "The DNA analyst may have calculated the statistics incorrectly for hundreds of other cases, but got the statistics correct for the case that the jury is hearing today.")

And then there is the botched science in the crime labs that leads to false negatives, so the case is never prosecuted. No blood, no semen, no drugs, equals no crime. The criminals remain in the street and the victims do not see justice. The crime lab may not even identify the false negatives because evidence is rarely re-analyzed a second time (because re-analysis is to proclaim that they botched it the first time.)

If anything, the judges should be pissed that accreditation auditors are not performing their job at finding (or reporting) problems in the crime labs. Judges should be pissed that so-called experts are not recognizing their scientific limitations, or that the so-called experts are failing to disclose "the whole truth." Judges should be pissed that the charlatans and the incompetent are wasting the time of the jury when adjudicated cases are found to be erroneous, sometimes purposely so, just to win.

If judges really wanted to make a mark, they should restrict those so-called experts from testifying in their courtrooms. Once the expert has been shown to be unethical, unscientific, or biased...ban them altogether.

Anonymous said...

What about "experts" that dont include DNA or fingerprints? So called scientific expert opinion. We had a SANE nurse who just happened to have had on a LIVE-STRONG bracelet as she testified. She used the bracelet to show how a hymen is pliable, can move, stretch and bend and still be completely intact, undamaged, no tears and no scarring even though this alleged victim had testified to being roughly molested over 200 times from the age of 11 thru 14. No evidence of any kind yet the jury decided better safe than sorry.

Anonymous said...

@Steven Seys-

The most famous fingerprint matching goof was that of Brandon Mayfield who was wrongly tied to the 2004 bombing in Madrid. The FBI described the fingerprint match as "100% verified". Oops.

Anonymous said...

You can't accuse them of something based on the fingerprints left at the crime scene. I don't even believe in DNA "evidence".

Anonymous said...

Care to elaborate?