CLPEX Chat Board

HPD says print unit fix won’t be cheap
City Council cool to $4 million funding request

http://www.chron.com/disp/story.mpl/met ... 53111.html

HPD says print unit fix won’t be cheap

City Council cool to $4 million funding request
By MOISES MENDOZA and JAMES PINKERTON
HOUSTON CHRONICLE
Dec. 4, 2009, 9:27PM

City Council members are balking at a Houston Police Department request for $4 million to keep consultants running its troubled fingerprint comparison unit, under fire after an audit accused it of botching analysis of fingerprints in criminal cases.

At a committee meeting Friday, some council members strongly criticized the HPD's plan to fix the unit, and suggested the entire department needs an overhaul.

“We've got a real problem here, and it's not just forensics … ,” said at-large Councilman Peter Brown. “We need reform in HPD, and that's what you all should be talking about.”

Police suggested using funds from city land sales, but some council members said that money should go to pay down the city's debt.

The meeting offered a prelude to next week when the City Council is expected to consider the funding request.

An audit of the fingerprint comparison unit, completed in October, found prints had been marked as unable to be analyzed or were overlooked completely in more than half of a random sampling of cases by the auditors.

Authorities are reviewing nearly 5,000 violent criminal cases from the past six years to see if they find similar problems in those cases. There is also a backlog of more than 6,000 fingerprint cases that have never been reviewed, covering the last two years.

As criticism of the department intensifies, Police Chief Harold Hurtt strongly defended his command staff's performance to a reporteron Friday, pointing out that the unit's problems likely were happening long before he became chief in 2004.

Criminal investigation
“As we go through the organization and we find these things, the messenger gets punished,” Hurtt said. “And this has been going on for many, many years in the organization and through many administrations.”

Hurtt added, “I'm not too sure that other folks in previous years didn't have the same knowledge but didn't have the courage, or no one had the commitment, to correct it.”

Three employees in the fingerprint comparison unit have been put on administrative leave, one part-time employee resigned under pressure, and a criminal investigation was launched into how one of the employees handled a piece of evidence at the conclusion of the audit, a police official said.

While no wrong identifications of suspects are known to have occurred, criminals may have gone free because analysts missed or didn't properly analyze fingerprints on evidence, officials said.

Debate next week is likely to be as harsh as it was Friday, when some council members compared the fingerprint unit's problems to botched DNA analysis that was discovered at the department's crime lab seven years ago.

Councilwoman Jolanda Jones complained that police have refused to release the actual audit report that's prompted their request for funds. Councilman Edward Gonzalez fretted that wrongful identifications of suspects may emerge.

Tim Oettmeier, an executive assistant chief who spoke before the council, said the department was “embarrassed” and added, “We clearly want to regain the public's trust.”

Oettmeier said the audit itself can't be released because of the pending investigation into the employees on leave.

Can't wait, HPD says
Under the plan he proposed, the $4 million would go toward paying for about 15 consultants to reduce the backlog, review past violent crime cases and rebuild the unit's protocols over the next two years.

Later the department will request more money to hire and train new staffers to replace the consultants. That could cost an additional $1.5 million to $2 million annually, Oettmeier said. Eventually the department wants to get its fingerprint comparison unit accredited by a national organization.

Council member Toni Lawrence suggested that the request for money could be tabled until January.

But Oettmeier said the department has the cash to fund the consultants only through the end of next month.

“The situation is so dysfunctional right now,” he said.

moises.mendoza@chron.com

james.pinkerton@chron.com

While no wrong identifications of suspects are known to have occurred, criminals may have gone free because analysts missed or didn't properly analyze fingerprints on evidence, officials said.

Can anyone provide more detail on what the discrepancies were exactly? I thought I read in another report that RS&A conducted the audit. I'm just curious as to what mistakes were made.

No doubt someone will expand on this, but I had it from another source in an attribution that I assumed was from this affair that it was substantially a matter of impressions deemed insufficient for examination and some deemed not to be fingerprints at all that were found to be so.

I tried another google news search and cannot find any "news" since that last article. I have heard from others via email that there were no erroneous identifications, but there were identifiable latents ruled by some HPD examiners as "no value" and there were apparently idents that were overlooked. We all occasionally fail to "see" one we should have identified, but the situation at HPD is said to have gone substantially beyond that. Other than that, I don't know much except what is reported in the first article. I don't know whether the root of the problem was poor training, overworked examiners trying to crank out too much work, ethical lapses by examiners, or what.

My purpose in posting that article was to provide a tool for examiners fighting the budget battle in these tough times. Taking short cuts or crimping on expenses up front not only leads to the perception of poor work, it can be much more expensive in the end. I think the message to the City Council is, "Pay now to fund the latent print unit properly, or pay later to clean up the mess."

The report can be found here, get it while it lasts: http://dig.abclocal.go.com/ktrk/AuditLatentPrint.pdf

I got it from the cached web page of a news story on a website which has since been removed.

http://74.125.155.132/search?q=cache%3A ... tartPage=1

HOUSTON (KTRK) -- We're getting a look inside the Houston Police Department's latent fingerprint lab. This comes hours after the department released a scathing audit outlining a number of mistakes.


Related Content
Story: Costly cleanup after HPD fingerprint lab mess

Story: Mistakes at HPD lab leads to review of cases
[ READ IT: Results of Latent Print Lab Audit]

Last week, the department fired three veteran workers and Tuesday released an audit which showed a number of mistakes that were made.


HPD says it will review print evidence from all violent crime cases over the past six years. That's roughly 4,000 cases. The audit released Tuesday focused on 156 cases tested by three workers, two of whom were fired.

The audit found mistakes in 58 percent of the samples from one of the workers. The other two employees saw 46 percent worth of mistakes each.

Police Chief Harold Hurtt says there could still be criminal charges as a result of the investigation.


(Copyright ©2009 KTRK-TV/DT. All Rights Reserved.)

$4 million to fix the problem? That's ridiculous.

Thanks for the post, Boyd.

I think that money was supposed to be spread out over a 2 year period and used to knock down the 6.000 case backlog, review 4,200 violent cases of the past 6 years. Keep the current work load current, develop SOP's and personnel to eventually move the ID Section to accreditation.

I am not sure how many consultants it would take to do that over a 2 year period. Just knocking down a 6.000 case backlog would seem expensive.

My 2 Cents: When administrators let a section go down the tube and then act surprised on how much it will cost to get it back to where it should be is something you would think they would teach in administrators school.

I think the cost shows that a lot of people and resources are needed to do the amount of work needed in a city this size. It's mind boggling how 3 examiners managed to keep up as well as they did.

Although I don't know anything about this agency or these examiners, I do know several other people and agencies that have had similar audits. Even though we try to represent situations numerically, I don't think the numbers always represent what people think they are representing. Here's an example, lets suppose that we say that on page 1 of this report it says that 59 cases were looked at and on page 3 it says that 58 cases were looked. One error on one report could be considered a 100% error rate. Or maybe this is 1 error on 2 reports (there are 2 dated reports here), which would be a 50% error rate. Of course, either seems unacceptable and I doubt that the person who wrote this report is under investigation right now. This is because we know that errors have to be looked at by their seriousness and not by a number.

In applied sciences, practicality and resources are considered when deciding what needs to be done, requirements differ based on the situation. We wouldn't expect a doctor to run the same expensive tests for a minor illnesses that we would for a major illness. There are different structural requirements for a 10-story building than for a 1-story building. In this audit, it would be interesting to know if the latents deemed by the consultants to be 'of value' were in property crimes or major crimes, but it doesn't appear that this kind of information was considered.

I’m also curious at the number of cases these examiners were completing every month, it’s possible that these examiners deserve a raise not an investigation.

I hope these examiners enjoy there time off during the holidays and I wish them all well!
Michele

Michele wrote:In applied sciences, practicality and resources are considered when deciding what needs to be done, requirements differ based on the situation.

Hi Michele

To expand on this thought of yours, I whole heartedly agree. When faced with growing backlog, I have always felt that working 40 cases per month, each 95% complete, was better than working 20 cases per month, each 99% complete. And either of those options is better than working 10 cases per month 100% thorough and complete. IMHO, "The greatest good for the largest number of cases" (quality and quantity) is a better philosophy than "Thorough and Complete in all cases."

Sometimes, one method of latent print development fast and dirty to recover 95% of the latents makes more sense than sequential processing to get a few extra latents while spending three times as long. And fast comparisons to grab one or two quick idents in a misdemeanor case makes sense so you can move on to a violent crime examination.

I don't know much about Houston PD and I don't know these examiners personally. If an agency is all about style over substance, you'll get the "100% effort on every case," without regard to the seriousness of the case or the backlog of the individual examiner. And the administration will use ASCLD-LAB to hammer the examiners. I, too, would be interested in knowing more about the backlog per examiner, and their duties (processing and comparisons, comparisons only, etc.). In a perfect world, every case would be serious, examiners would work a reasonable number of cases per month, could work everything thoroughly, and always get the right answer (ident or exclusion). We should strive for it, even though the world isn't perfect...yet. I only hope that the examiners (the "Babies" if you will) don't get thrown out with the bath water.

I don't post much here but I feel I should post on this subject. I agree with you Michele. the first crime lab director I ever worked for told me something that has stuck with me through the decade I have been involved in forensics. He said "Forensic science is one of the only careers that DEMAND perfection at an impossible level". I did not really understand that statement when I first heard it...I had only been in the field for about two weeks at that point. Now I fully understand it.

It would almost seem that a forensic analyst, at some level, is doomed to fail from the start. Honestly, if a company decided to audit anyone's lifetime of casework (no matter the sub-field) errors would be found at some level (I, II or III). That being said I think the forensic community should take a long hard work in relation to how we handle errors from practitioners. Do we require perfection at an impossible level? If an analyst makes a mistake, no matter the error type, does that really negate the education, knowledge and experience that they have accumulated over their career? Does it make them a "bad" examiner or just a "human" examiner?

It is my personal belief that this a QA/QC issue. With a good QA/QC policy in place that requires both internal and external annual review of active cases, blind proficiency testing and bi-yearly audits alot of these errors will be caught early and can be addressed. Unfortunately, the way this audit has been handled by administration and the media will more than likely end the career of these examiners or greatly impact them and I cannot help but wonder if this is more of a ploy by the agency to get 4 mil. rather than actually correct the problems that exist while maintaining their existing personnel. I applaud the crime lab for auditing the unit and I applaud the auditors for doing a thorough job of the audit...I just think that firing the examiners may have been a stretch.

Of course, this is simply my personal opinion and I don't know all of the facts behind the errors that were encountered or any underlying issues that may have existed to prompt this response to the audit. In the end it may be completely justified that the HPD made the move they did. Especially with the quote from the chief stating that "criminal charges may result'. That says to me that their may be more to that story than what we are hearing?

ashepherd wrote:It would almost seem that a forensic analyst, at some level, is doomed to fail from the start. Honestly, if a company decided to audit anyone's lifetime of casework (no matter the sub-field) errors would be found at some level (I, II or III). That being said I think the forensic community should take a long hard work in relation to how we handle errors from practitioners. Do we require perfection at an impossible level? If an analyst makes a mistake, no matter the error type, does that really negate the education, knowledge and experience that they have accumulated over their career? Does it make them a "bad" examiner or just a "human" examiner?

I've long held that the notion that we do do not tolerate error and that an examiner is finished after a revealed official error is a warm but empty feeling. It often cited as evidence, along with rigorous standard process, that error, careless or otherwise, carries such severe consequences that the consequence of error is a powerful factor in the rarity of error. At the same time, it's really not a matter of error never being tolerated, in terms of the examiner continuing to practice. My impression is that if the examiner's proficiency is implied by the sum of their training, performance, and position, an officially reported error will not end a career, while another examiner whose error can be blamed on deficiencies in training or adherence to standards is finished. So in a very real way, we do often apply the very rational approach that a fully proficient and conscientious examiner may produce error and that the examiner need not be lost but can be brought back to an expectation of no future error through various corrective measures. And we also recognize examiners whose training and self-determination to good practice is so deficient that they essentially would have to start over from the beginning before they would ever practice again. We just don't have a process to do any of this consistently. And our more or less quasi-official attitude about what sorts of error we get excited about isn't at all well defined.

The problem is that the sanctions can be invoked arbitrarily. And the pretense to zero-tolerance that we imagine self-disciplines practitioners can be cited by others to imply that examiners are ultimately responsible for all problems. Whether error is handled constructively then depends on how the employer wishes to interpret the failure. The FBI response to Mayfield was probably about as good as it's likely to get. While I'm not quite convinced that every factor that might have been at work was revealed, the corrective measures address everything known and probably anything else that might reasonably be speculated. It became a lesson learned and applied. I think that's one of the benefits of a having a true OIG available. You don't get to assign or imply that error was just a function of how examiners did the job. You learn how to get better in a meaningful way that goes beyond reinforcing the fear of examiner sanctions, because most OIG's root out causes along with effects. Consultants can only address the mission dictated by the employer, in this case a "limited scope technical performance review" directed solely at the analysts. One might usefully contrast this with the same consultant's report on the Boston P.D. unit. We learned a lot from the Mayfield OIG report and the Boston report.

In the Houston inquiry, there is no analysis of why three examiners assigned to latent print examination erroneously rejected potentially valuable evidence in a high percentage of their cases. The numbers are astounding, made more significant by the fact that the kinds of error identified are endemic in the unit. That inescapably implies specific identifiable problems. The report does establish through examiner interviews the de facto standard for sufficiency for further analysis. What was not asked (being outside the "limited scope") was why the standard was violated. It seems to me to substantially defy belief that three examiners whom the consultants assumed had been trained to competency, one of whom was a CLPE, could not recognize Galton points present in a latent or couldn't count to seven. That they very likely could in no way points to one or another source of the errors as reported. It does not wholly absolve them, nor does it wholly condemn them. Neither does it absolve anyone else. Without understanding what happened, we have nothing to take away from it that helps us in any way avoid the effects of those mistakes elsewhere.

What remains to be seen is whether the reasons will ever be revealed or the administration or the consultants (assuming the funds are provided) will come in and establish a system to eliminate the reasons without articulating them.

The fact I agree with Michele is not surprising, and ashepherd, you are in with good company. Pat, I agree with your premise, but I think certain policies and procedures need to be administrative decisions, not technical matters that rest solely with the examiner. I have done enough reviews, audits and inspections to know that any expectation of complete perfection is foolhardy. At the same time, I have observed when a horrific backlog has become an excuse for lowering standards until standards no longer exist. My read of the audit is that it became easier and easier to determine a latent print was of "no value" until good evidence was routinely tossed aside. Huge backlogs are demoralizing, but they have also become a fact of life for far too many agencies.

Where the line is between accomplishing the greatest good for the greatest number and doing sloppy, second rate work is difficult to determine. Pat, I think you have a pretty good moral compass, but not everyone does. That's why I think the administration has to state clear expectations. Few do, unfortunately. But one closing comment from a regular to this board is the key - take responsibility for your own actions. I have a hunch from the consequences mentioned that a whole lot of rationalizing and weak excuses went on but not a lot of responsibility.

I do get disturbed by the panacea for technical challenges is to become accredited. No accrediting body comes in to proclaim nothing but good work is going on. Accrediting agencies just note that polcies and procedures are in place to promote good work, things such as an effecitve QA program and training records that document ongoing training or lack of it. The primary purpose of checking casework during accreditation inspections is to ensure established procedures meet general acceptance and are followed by all.

Forensic analysis of this sort is an easy target. There's a nice bright line of performance measurement in individual cases, like the sufficient to examiner standard they applied here. You can't do that kind of audit on, for instance, investigations. There's no minimum standard of rigor. I can always do something more in an open investigation. Stir the street more. Canvas farther out. Check pawns in the next county. Go stir the suspects again. Etc. But few can go all out on every case, so the priority goes to the majors and the apparent clearables. Backlog management was easy. No movement in 30 days, close it. Minimum was contact the victim, talk to the witnesses, and jack the suspects. Easy to do and easy to review. Obviously, some detectives did all they could with the time allowed, and others did as little as they could, closing their cases with less effort in the same 30 days. But all that was deliberate and open management decision making. It was response to backlog or potential backlog. And it was management's responsibility to also monitor who was working appropriately. Backlog has to stimulate a management response. No particular management way of handling backlog is necessarily wrong. The exception, the only really wrong way, is to demand the backlog be kept low without making it possible, because you also demand completeness. That forces either some kind of showdown in which the lowest rank loses or fudging the work and hoping no one notices. (In the investigative example, the fudging was real hard to detect.) Nasty position either way. It's hard to say how much responsibility for something like this you can reasonably expect from an analyst under impossible conditions when he has larger responsibilities. But it's management's job to see that the analyst never has to make that choice. And it's the job of the administrator above that manager to see that the manager, if necessary, is allowed to pick a lesser doable option than perfection, and it's that administrator's responsibility to carry that hard necessity up the line and the chief's ultimate responsibility to see that he doesn't get killed for bearing the bad news.

I used to tell my CJ management students that the reason companies put "MANAGER" on the manager's office door was because otherwise, you couldn't tell they were working. If you were walking down the hall of the police agency and took someone past an unmarked door and they asked, "What's he doing?", it wold be embarrassing to answer, "Well, uh. He's - uh - captaining." So you put the title on the door. You can only measure how well he's captaining by the quality of his people's work. THAT's the captain's product. For a long time, the erroneous product in that unit has also been the product of someone with his title on his office door. The numbers are his, too.

Interesting blog on the subject.
Kudos Jennifer for Boston, though (once again)
http://www.crimelabreport.com/library/m ... 2-2009.htm