A Review of the FBI's Handling of the Brandon Mayfield Case #4 application of the 1 discrepancy rule

As every examiner knows, latent and exemplar images are never exactly the same. The reasons are many; how the exemplar was recorded or where the latent was deposited, etc. The experienced examiner during the analysis and evaluation must be able to understand and explain any discrepant points found in either the latent or examplar images. In this case we had experienced examiners determine the discrepant points of identification were the result of latent distortion, probably caused by the bag. The examiners made a subjective determination which was incorrect. The panel noted theFBI examiners failed to follow the 1 discrepancy rule , if there is 1 unexplainable discrepancy in either the latent or exemplar image an individualization cannot be made. In this case the examiners did re-evaluate the prints and tried to get the Spanish examiners to agree with the FBI examiners individualization, I don't think they ignored the rule , they felt they could explain the discrepancies. The examiners would have given the Spanish examiners their interpretation of reasons for he discrepant points.

The thing to remember is if highly skilled and experienced examiners, which each of these examiners is can make this type of mistake, what about the less skilled and experienced examiners. Most individualization's are done by examiners with much less experience and training than the FBI examiners. Currently the most effective prevention is for the defense or court to have the fingerprint evidence re-evaluated by an unbiased 3r party.

Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

A Review of the FBI's Handling of the Brandon Mayfield Case;Problem #3 Faulty reliance on Level III Detail

In the Mayfield case the FBI examiners increased the total minutiae they felt they had from 10 points of Level II detail to 15 total points utilizing level III detail. Level III detail can include a number of things from ridge shape and pores to scars and this information can be beneficial in individualization when properly used. It would appear that they continued the " circular reasoning " utilizing the level III detail. The examiners justified the the ten level II points based on the level III detail; ridge shape,pore placement and incipient dots. While I did not see the original images, from what I was able to see, the latent lacked clarity. On poor quality images it is very difficult to obtain reliable level III detail, the such small ridge characteristics are usually compromised by background clutter and distortion. The FBI had several different copies of Mayfields know prints and some of the level III detail was not present in all the copies, which should have caused the examiners concern. While level III detail can be helpful, it becomes less helpful as clarity decreases. I want to emphasize how important it is to utilizie multiple copies of an inked image if they are available when working poor quality images.

Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

A Review of the FBI's Handling of the Brandon Mayfield Case, unclassified executive summary. Reason #2

The 2nd. significant factor noted by the OIG was the examiners of some features adjusted or influenced by "backward " reasoning. In effect after determining 10 points of "unusual similarity" the examiners found additional features by working backward from Mayfields file print. The examiners found features in the exemplar and then looked for them in the latent image resulting in "murky or ambiguous details" being erroneously identified as points. When an examiner is dealing with a poor quality latent that lacks clarity, it is tempting to look at the exemplar and work backward but as is clear from the Mayfield case it can have disastrous results. This backward or circular reasoning allowed the examiners in Mayfield to report up to 17 points of identification. A jury or defense attorney presented with a case like Mayfield would be ill equipped to determine that an error had occur ed and then convince a jury of the error.

If you watch any of the shows that highlight current criminal trials ,you'll often hear the analyst talk about having 8 points in fingerprints as a standard, they indicate if there are 8 points the print is an identification. I would suspect many lawyers are under the same impression that if there are 8 points case closed, as we see in Mayfield this is not the case. I would also go back to the point I've made numerous times in the past; the examiners in the Mayfield case all had more training and experience than most examiners testifying in court today. Even highly experienced latent examiners with training can make an identification error, the only way to prevent the error from causing disastrous on sequences is to have the evidence evaluated by an experienced examiner.

Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

Additional Point on Item 1 of the FBI Report

One other point I would like to make so it's not lost in item 1 of the FBI report. The 10 of the points of identification finally used to identify Daoud,were also used by the FBI to incorrectly identify Mayfield. Why is this significant? most cases of fingerprint identification accepted in court today are brought in by latent examiners with less experience and training than the FBI examiners. With no "systematic study of the rarity" of finding such a constellation of similar points of identification, how can we expect less experienced examiners to make identifications on similar difficult identifications.

Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

A Review of the FBI's Handling of the Brandon Mayfield Case, unclassified executive summary.

Went back over the executive summary on the Mayfield error back in 2004 and want to discuss what the Office Inspector General determined were the causes of the error as well as what the OIG suggested to prevent similar errors in the future.
The report listed six (6) causes for the error and in this blog I'd like to discuss the 1st. major contributing factor in the error. The OIG found the primary cause of the error was the similarity in Mayfield's print with the latent found in Spain. " Despite the unusual similarity in the relationship between points on the Mayfield and Daoud prints, Mayfield and Daoud did not have identical fingerprints." It also pointed out there are no studies on how often situations similar to this occur but anecdotal reports this as a rare occurrence. The identification was an IAFIS suspect and as the OIG pointed out "The enormous size of the IAFIS database and power of the IAFIS program can find a confusingly similar print." The OIG in my opinion was correct in highlighting this as the major factor in the error. As databases increase and algorithms improve there will be more suspects produced, that will have unusual similarity between points but will not be the individuals print.
So can this type of error be prevented ? There will continue to be similar errors we can minimize the numbers with training and utilization of fingerprint experts for the defense. In the Mayfield case you had four (4) experienced and well trained fingerprint examiners that made this bad identification.


Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

The waiting game

The following article brought out a number of good points;
http://www.suburbanchicagonews.com/beaconnews/news/1781113,2_1_AU21_PRINTS_S1-090921.article
"And of course, he wants the assailant caught as soon as possible, which is why the seven-month wait for a fingerprint match has him upset. Police do have a match on the suspect now, he said, but now they're asking his relative to identify someone she hasn't seen in seven months. " A backlog causing a delay like this not only increases the likelihood of additional criminal activity by the criminal but also the possibility of an escalation in the violence in future crimes.
The other point made by this article: "Aurora's crime lab doesn't: an Automated Fingerprint Identification System, or AFIS. According to Leroy Keith, an Aurora alderman who also works as assistant director of DuPage County's crime lab, the equipment itself costs about $50,000. But that's not the big expense: You also need people trained to make the final matches." As indicated in earlier blogs having trained latent fingerprint examiners requires 2 yr's minimum of full time training to become a competent examiner. The expense and time can often tempt smaller agencies to not fully commit to the required training increasing the possibility for error.
While technology continues to improve we should expend similar resources to insure the individuals working with the automated systems are the highly trained individuals required for the systems.
Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.

Another Erroneous Fingerprint Identification

Posted an article in a number of places a few days ago from the NY Daily News about a large settlement for an individual incarcerated for 523 days because of an error in a fingerprint comparison. The thing that caught my eye 1st. was the liability cost,which while significant was not exorbitant.

Then I looked at the article again to see the real cause of the error, because unlike the FBI Mayfield error, this was just a robbery and in NYC that is not a high profile case.

So what went right and what went wrong on this fingerprint comparison. What went right was an experienced examiner who was confident enough in his skill sets said " you know what ? this is a screw; this is not his fingerprints." If it were not for this detective the individual in this case would have been convicted and gone to prison. What went wrong ? The original comparison followed the ACE-V individualization model followed by latent examiners, in that you had 2 experienced latent examiners incorrectly identify the individual. The legal aid attorney advised the victim to take a plea agreement that would have been 5 yr's, an indication that the fingerprint evidence was not re-evaluated by the defense.
So what could have been done to prevent this from happening ? As I've stated in previous blogs having fingerprint evidence evaluated by the defense is key to preventing miscarriages of justice. It is really the defense attorney who is responsible to insure the evidence is valid in the adversarial system of justice we have in the United States. I know that resources for expert testimony are limited especially for legal aid but the cost of a fingerprint evaluation should significantly less than expert testimony so it should not cost more than a service call from an electrician. The images can be sent to the expert digitally via Internet or photographs via snail mail for an initial evaluation.

Incorrect fingerprint comparisons do occur and with the automated systems producing many more suspects having strong similarities the numbers of error most likely will continue to increase. The best option currently available is for the defense to have the fingerprints evaluated by an unbiased 3rd. party.

Bob McAuley
Dir. Operations/Training
Forensic Biometric Identification Solutions LLC.