Sunday, October 17, 2010

Top Ten List: Forensic Sciences That Aren't Science

At the recent Texas Forensics Science Seminar sponsored by the Criminal Justice Integrity Unit of the Court of Criminal Appeals, I attended presentations by DPS crime lab experts on ballistics, toolmarks, hair (and other trace) evidence, and fingerprint examination. Much of the discussion at the event was framed by the 2009 report (pdf) from the National Academy of Sciences which was particularly hard on forensic disciplines which rely upon subjective comparisons, from fingerprints to hair comparisons to shoeprints. After hearing DPS' own experts discuss how they perform these tasks, it's difficult not to agree that the basis for these practices is not scientific in the same sense as, say, DNA analysis.

Dr. Jay Siegel, who was one of 17 people on the NAS panel that produced the report, told the seminar audience to closely examine terminology from experts providing so-called "pattern evidence": Ask them, what is the definition a "match"? By what methodology can the examiner individualize their findings? And beware "weasel words" like "consistent with," he said, or statements that an item "could not be eliminated as the source." Later in the day, however, the ballistics/toolmarks expert who presented on those subjects for DPS admitted he used those phrases in his work and defended doing so. (This despite the fact that even bullets fired from the same gun may have different markings, he said.) For these disciplines, the NAS' critique that they're subjective comparisons, not science, was impossible to avoid. That's true even (perhaps especially) regarding fingerprints, which for generations have been considered the gold standard of modern forensics. Crime lab experts are able to make these distinguishing judgments, we were told, based on their "training and experience," but when the process was described in detail, it all came down to "After looking at them closely, I think these two things look alike."  There's an argument to be made that that should be enough, that forensic testimony shouldn't "gild the lilly" by claiming experts can individualize their findings to a greater extent than is scientifically valid.

Even more subjective, and typically performed by poorly trained police officers in the field at DWI stops, is the Horizontal Gaze Nystagmus (HGN) test, where officers are told to hold a flashlight or other object at the edge of the driver's peripheral vision, move it horizontally back and forth, and look for minute cues regarding their eye movements that weren't even obvious on the training video. The DPS trainer who gave the seminar presentation readily granted that officers who didn't routinely perform the test and stay up on their training often performed the procedure incorrectly or misinterpreted the results, though he defended the test when performed properly. Either way, it couldn't be a more subjective technique, even when performed correctly. Another speaker from the American Academy of Forensic Sciences said in his opinion HGN shouldn't be considered "science," though courts routinely accept its results as probable cause in DWI cases.

When you add in all the subjective disciplines to other forensic methods with gaping methodological flaws, the list of forensic sciences that aren't really science grows quite long. The following top-ten list isn't comprehensive - just the first ten that came to mind - and I encourage readers to add to it in the comments.
Top Ten Forensic Sciences That Aren't Science
1. Fingerprint examination
2. Horizontal Gaze Nystagmus (field sobriety test)
3. Ballistics
4. Blood splatter
5. Hair/trace-evidence comparisons (non-DNA)
6. Toolmarks
7. Polygraphs
8. Arson (older cases)
9. Autopsies
10. Dog-scent lineups
I'm not arguing that these forensic disciplines are always wrong; in fact, subjective comparisons by a person (or perhaps even a dog) who performs the task frequently may usually turn out to be accurate. But that doesn't make what's going on scientific or really more than an educated opinion as opposed to a definitive, science-based finding of the type associated with DNA or identification of controlled substances. It seems unwise to overstate conclusions from those disciplines when their scientific basis has been so widely and publicly questioned and the needed research to validate them hasn't been done.


Anonymous said...

The problem with the forensic science Top 10 is not in the science. It is in the police officers, prosecuters, judges and juries who rely soley upon that forensic evidence when no other evidence, witnesses, or statements are present. In turn relying soley on forensics to find someone guilty or not guilty.

This is especially true with juries who are brianwashed by CSI and other TV police dramas based upon forensics to solve crimes in 1hour time slots.

The police got away from "good police or investigative work" and are relying to heavily upon forensics.

Prison Doc said...

Wow..glad to see I'm not the only one who knows that lateral gaze nystagmus is bullshit, just like the rest of the field sobriety test.

I've always been amazed that after a successful win in a jury trial the victorious party steps before the cameras and vows their "faith in the jury system". Not me...but unfortunately it's like Churchill said about democracy--"the worst system in the world except for all the others".

Anonymous said...

Under your logic, Grits, the field of medicine is not a "science" either, since most medical diagnostics is subjective.

Anonymous said...

Subjective expert opinions are actually very common in civil courts, especially in the areas of medical malpractice, product liability, and many personal injury cases. Accident "reconstructionists" routinely give opinions on liability or causation in vehicle accident cases. All of these "experts" give opinion testimony with varying degrees of scientific "certainty." Juries are quite capable of figuring out what's credible and what's not.

Jim B said...

Anon 4:31 said: "Juries are quite capable of figuring out what's credible and what's not."

I couldn't disagree more. This forensic science evidence is presented as scientifically valid to bolster it's credibility. As in Dep. Pikett's claim that his dogs were 100% accurate. How he got by with that for so long makes one wonder.

Juries want to hold the accused accountable for the crime. They do not wish to think the DA and Police would knowingly prosecute an innocent person. Increasing the value of "forensic evidence" by claiming some false, scientific basis is just plain wrong.

Anonymous said...

I know you can't learn anything from autopsies. They should stop performing them. Plus, there is no way to tell if someone is drunk.

Gritsforbreakfast said...

2:35, of course it was actually advances in medical science, specifically DNA, that highlighted flaws in traditional forensics. Apply medical rigor to forensic analysis and it won't stand up.

4:31, if it's just opinion testimony, say as much. Don't claim it's science when it's not.

6:23, you're full of red herrings today. Let me encourage you to go back on your meds.

Anonymous said...

Grits has no problem with expert long as it's offered by the defense. And he knows that courts can't preclude a defendant from putting some hocus pocus witch doctor on the stand without creating a point of error for appeal. But yet he'd love nothing more than to create a system where a criminal conviction is nearly impossible to obtain. There isn't a single kind of evidence, including confessions, that he hasn't complained about at one time or another.

Anonymous said...

Drug dog hits.

Once transported 400 keys of pure cocaine in an otherwise clean trunk of a car a short distance. About 12 to 16 hours later, I had a drug dog run on the trunk. Nothing. Nada. Zip.

Some handlers claim their dog can ID a vehicle, house or boat that previously transported narcotics; ask to see their log of all successful AND unsuccessful stops - all of 'em along with all training documentation (Where the unsuccessful "hits" should have been documented) otherwise how can you prove the capability of your animal?

Record keeping, I understand, is done with explosive-trained dogs but I've never heard of it for drug dogs. :~)

escalante blogger said...

That an investigation terms where police uses.

Anonymous said...

"There isn't a single kind of evidence, including confessions, that he hasn't complained about at one time or another."

He wants the thug back in your neighborhood no matter what. What a liberal!

Anonymous said...

I wondered if people would wake up to the fact that ballistic forensics isn't a science. I've been an amateur gunsmith most of my life, and I've always wondered how a prosecutor could make a case by proving that rifling marks on a bullet prove that a bullet came from a certain "unique" gun. The reality is that in firearm manufacturing, the same chamber reamer and/or barrel broach are used to manufacture several barrels before those tools get too dull and are discarded for sharper replacements. However, before they are discarded, it wouldn't be unusual at all for them to create guns that left identical marks on bullets. To that, you might ask what are the chances that these "identical firearms" would exist in the same community, or even the same state. To that I would reply that over the years, I personally owned two revolvers and two rifles that were one serial number apart from each other, and most likely, were manufactured with the same reamers and broaches. I would be willing to bet that these pistols or rifles would create identical rifling marks on test-fired bullets.

Petra de Jong said...

That HGN stuff is ridiculous. Don't cops carry breathalizers with them over yonder? Over here, they do. No subjectivity involved. And comparing fingerprints by eyesight??? We've got computers do that...

Anonymous said...

There are numerous consecutively manufactured barrel and breach studies in the field of forensic firearms identification that prove individuality between components consectutivly made by the same tool. Just because one is unaware of the studies only reflects ignorance of the field, not the implied deficiencies of biased journalists and uninformed public. The authors of the NAS report were mislead by uninformed "experts". Read a book.

Don said...

Petra D Roadside breathalyzers? Maybe there's no subjectivity but there is junk science. And notorious inaccuracy.

Anonymous said...

Yes, the NAS clearly relied on uninformed experts. Those experts included Joseph DiZinno, Asst Director Lab Div. of the FBI, Jan Johnson, Lab Dir. or IL State Police, John Morgan, Dep. Dir. for Science & Technology NIJ/DOJ, Bruce Goldberger, President Elect American Academy of Forensic Science, Bill Marbaker, President American Society of Crime Lab Directors, Peter Striupaitis, Chair IAI Firearm/Toolmark Committee, etc. They only brought in people from the US Secret Service, the FBI, ATF, various state police agencies, forensic science programs at universities, SWGGUN, IAI, etc.

Clearly DPS knows more about what they do than these "experts" that train people in the field, certify people in the field, have the biggest labs in the field. And those articles, reports, and studies they cite from things like the National Research Council are just bogus compared to the guy with a chip on his soldier at DPS who wants to keep his job and does not want to be questioned about whether what he does is or is not science.

Anon 9:16 hit that one right on the head.

Larry said...

I’m a bit familiar with the natural sciences and what constitutes scientific evidence being that I hold a M.Sc. in Physics, with a thesis in Observational Astronomy.

What passes in Texas Courts as scientific evidences is a ridiculous joke. I NEVER get selected for jury service. My friends from grad school never get picked either. It would seem that all a DA, or ADA has to see on your jury form is your occupation listed as Physicist, and you’re passed over.

Gritsforbreakfast said...

@8:56/7:14 - "There isn't a single kind of evidence, including confessions, that he hasn't complained about at one time or another." ... "What a liberal!"

FWIW, the Uniform Code of Military Justice has long required corroboration for confessions to secure a conviction. Does that make the US military "liberal" because they require a greater burden of proof than Texas? Just curious.

7:49, good point on drug dog hits as probable cause. Their work is mostly undocumented and sporadically tested, by all accounts.

9:16, from what was said at the seminar, some of the studies you allude to didn't utilize "blind" testing or had other methodological problems. As 10:02 points out, the NAS group heard from plenty of experts and assessed the research you allude to and more, they just came to a different conclusion. Adding to their credibility, unlike anonymous blog commenters, they were willing to cite their sources in a lengthy report and publicly sign their names to their opinions. You tell me to "read a book," but the problem in this conversation is that you need to read the NAS report. Either you simply haven't done so or you've chosen to lash out in reaction rather than accept the challenges it presents on their face.

Soronel Haetir said...


One discipline I'm surprised didn't make your list is bite mark evidence (especially when said bites are in poor preservation media such as skin).

Anonymous said...

The NAS report is a hardcover book and I bought and read it. I did not steal a .pdf file of the report. Most journalists stop at the end of the 33 page summary or they get a book report of the summary at some seminar. The fact is, sufficient information was supplied to the academy and they cherry picked journal articles and studies for their own agenda. Plus, it was committee that met a few times. Not all outside experts were given equal time and some were ignored. Money is needed for the field to strengthen the known theories and facts and the junk science needs to be deleted. Thats why "strengthening" is in the title.

Anonymous said...

Any examination in which a human is involved is going to be subjective, even DNA or Drug Chemistry. An analyst is still needed to interpret the results, therefore conclusions are always subjective based on objective information. This doesn't make them more or less valid.

Anonymous said...

Is psychiatry a science? Don't they go with whoever is paying them? My little sister was like that.

Anonymous said...

The use of fingerprints to identify people has never been proved to be valid. It is used to wrongly convict.

Anonymous said...

Sorry, 10/18/2010 09:16:00 AM, I didn't read the book. I actually build and test-fire firearms and components. Sure, there could be some very minute differences in tool patterns, even from the same tool. However, there is a much increased chance that their won't be any such difference. A chamber reamer will score virtually the same pattern against the boot of the riflings until the reamer itself starts to dull beyond use. And, that can take several barrels or more. I have reamers that I have used many times over the years on many barrels, and they are still usable. The reamer is much harder material than the barrel steel. Same goes for a barrel broach. Also, depending upon how bad a barrel is fouled before test-firing, that can also leave a unique pattern on a bullet. This pattern can completely change depending upon whether or not the barrel is clean, and how many times it is successively fired. Lastly, as a bullet is fired, not all of the copper jacket stays intact as it travels the length of the barrel. Small bits of copper are galled off, and can re-weld themselves further back on the bullet jacket, causing inconsistent rifling marks. Some of these copper galls can remain in the barrel, and be picked up by the next bullet fired, causing inconsistent rifling marks.

In short, I think there are enough variables at play that, in many cases, prevent one from determining conclusively that a certain individual firearm fired a certain bullet. Certainly, they would raise enough doubt to disqualify this type of ballistics testing as an exact "science."

Of course, as you say, I didn't read the book.

Anonymous said...

You really got the people who rely on this pseudo-scientific horsecrap for a living up in arms with this post. Well done.

abster said...

I am a huge critic of forensic pseudoscience, and a big fan of the REAL science -DNA, for example or Mass Spectometry/Gas Chromatography, etc. Or properly calibrated breath test instruments. As for HGN, it is VERY dependent on individual skill, whether performed by a cop or an MD or DO. I' e personally refuted another officer's HGN reading, for example. However, if you are good, it's very reliable. I can say with the case facts to back me up, that i never tested somebody with 4 cues of nystagmus who did not blow a .09 or above. Out of over 80 such incidents, and i never found 6 cues where the person blew below a .12. That being said, if suspected of DUI, you have no way of knowing how skilled the officer is. Imnsho, there should be rigorous testing of officers with let's say 50 test subjects at various levels of impairment, done double blind, and no officer should be allowed to use or testify to it without yearly certification. If I was a DUI suspect, i'd have complete confidence in the breath test and blood test, and in my state you have the right to demand a second blood test by a doctor of your choosing, so you can exercise that and get your own lab test. But I'd say no ro the HGN unless I knew the officer was competent

Great article. Any changes that help clear the innocent as well as finger the guilty are a good thing. Junk science is an abomination