Nov P12-23 Nov Fea 10.17

Tarnish on the ‘Gold Standard:’ Understanding Recent Problems in Forensic DNA Testing

• In Virginia, post-conviction DNA testing in the high-pro- ﬁle case of Earl Washington, Jr. (who was falsely convicted of capital murder and came within hours of execution) contradicted DNA tests on the same samples performed earlier by the State Division of Forensic Sciences. An outside investigation concluded that the state lab had botched the analysis of the case, failing to follow proper procedures and misinterpreting its own test results. The outside inves- tigators called for, and the governor ordered, a broader investigation of the lab to determine whether these problems are endemic. Problematic test procedures and mis- leading testimony have also come to light in two additional capital cases handled by the state lab. 2

• In 2004, an investigation by the Seattle Post-Intelligencer documented 23 DNA testing errors in serious criminal cases evidence has long been called “the gold standard” handled by the Washington State Patrol laboratory.3 DNAof forensic science. Most people believe it is virtually infallible — that it either produces the right result or no • In North Carolina, the Winston-Salem Journal recently result. But this belief is difficult to square with recent news published a series of articles documenting numerous DNA stories about errors in DNA testing. An testing errors by the North Carolina extraordinary number of problems related State Bureau of Investigation.4 to forensic DNA evidence have recently come to light. Consider, for example, the • The Illinois State Police recently can- following: celled a contract with Bode Technology Group, one of the largest • The Houston Police Department (HPD) independent DNA labs in the country, shut down the DNA and serology section expressing “outrage” over poor quality of its crime laboratory in early 2003 after a work.5 television expose revealed serious deficiencies in the lab’s procedures, deficiencies • LabCorp, another large independent that were confirmed by subsequent investi- lab has recently been accused of botch- gations. Two men who were falsely incrim- ing DNA paternity tests.6 inated by botched lab work have been released after subsequent DNA testing While these scandals are bad proved their innocence. In dozens of cases, enough, the problems with DNA evi- DNA retests by independent laboratories dence do not end there. A close look at have failed to confirm the conclusions of the field shows that DNA testing errors the HPD lab. The DNA lab remains closed have been popping up all over the coun- while an outside investigation continues.1 try. Many of the mistakes arise from By William C. Thompson

10 WWW.NACDL.ORG THE CHAMPION cross-contamination or mislabeling of While there have always been bad error in DNA testing is so low as to be DNA samples. Problems of this type have labs, their shoddy work has been difﬁ- negligible, but growing evidence sug- been documented in Minnesota,7 North cult to detect because the worst labs tend gests otherwise. Carolina,8 Pennsylvania,9 Nevada,10 and to be found in jurisdictions that have An important source of evidence on California.11 Tellingly, one of the private historically shielded crime labs from the nature and frequency of these prob- labs hired to retest DNA evidence in cases external scrutiny. For example, it is now lems is “contamination logs” and “cor- that were botched by the Houston Police recognized that the Houston Police rective action ﬁles” that are maintained Department Crime Lab has itself pro- Department (HPD) Crime Laboratory by some DNA laboratories. Under a

duced false matches due to sample mix- did grossly inadequate, incompetent and guideline issued by the FBI’s DNA U 12

ups. A particularly ominous sign of biased DNA and serology work for well Advisory Board in 1998, forensic DNA N underlying problems is that accidental over a decade before a team of television laboratories are required to “follow pro- D transfers of DNA among samples from journalists exposed the problems in late cedures for corrective action whenever 18 E

different cases being processed by the 2002. Defense lawyers did not (and proﬁciency testing discrepancies and/or R

same laboratory have produced several probably could not) expose the lab’s casework errors are detected” and “shall S 13 false “cold hits.” problems because Harris County maintain documentation for the correc- T While most of the problems are due (Houston) judges routinely denied tive action.”19 While many laboratories A to inadvertent mistakes, a number of requests for discovery of underlying lab- ignored this guideline, some laboratories N cases involving dishonesty have also come oratory notes and for expert assistance (probably the better ones) have begun to D to light. DNA analysts have recently been in evaluating DNA evidence. Indeed, keep records of instances in which, for I fired for scientific misconduct, and under a policy of the Harris County example, samples are mixed up or DNA N specifically for falsification of test results, District Attorney’s Office, that defense from one sample is accidentally trans- G by a number of forensic laboratories, lawyers rarely challenged, the defendant ferred to another samples. R including labs operated by the FBI,14 could not even get a copy of laboratory The surprise for defense lawyers E

Orchid-Cellmark (another large private reports in his case until the trial began. who have managed to gain access to these C 15

DNA laboratory), the Ofﬁce of the Crime labs in Virginia and North ﬁles is how voluminous they are. Errors E

Chief Medical Examiner in New York Carolina have also received little scrutiny occur regularly. Files from Orchid- N 16 17 City, and the United States Army. In all in the justice system due to severe limi- Cellmark’s Germantown, Maryland T of these cases, the analysts were caught tation in those states on the availability facility, for example, show dozens of faking the results of control samples of discovery, funding for independent instances in which samples were contam- P designed to detect instances in which experts, and funding for indigent inated with foreign DNA or DNA was R cross-contamination of DNA samples has defense in general. somehow transferred from one sample O occurred. But these problems could not to another during testing. I recently B L

So what is going on with DNA test- remain hidden forever. Journalists have reviewed the corrective action ﬁle for an E

ing? How can we explain this sudden rash played a big role. In Houston, the prob- accredited California laboratory operat- M of problems with “the gold standard” of lems were ﬁrst exposed in a series of ed by the District Attorney’s Ofﬁce of S

forensic science? How can a test that has exposes by television reporters who were Kern County (Bakersfield). Although long been advertised as virtually infallible assisted by academic experts. Excellent this is a relatively small laboratory that I produce so many errors? And what is investigative journalism also helped processes a low volume of samples N

behind the recent spate of dishonesty expose problems in Washington State, (probably fewer than 1,000 per year), F among DNA analysts? The answers to Virginia and North Carolina. Another during an 18-month period, it docu- O

these questions are, in my view, intercon- important factor has been post-convic- mented multiple instances in which R

nected. Some serious underlying prob- tion DNA testing. A number of errors (blank) control samples were positive for E lems with DNA testing that have existed have come to light because post-convic- DNA, an instance in which a mother’s N for a long time are beginning to come to tion DNA tests contradicted tests per- reference sample was contaminated with S light. What we are seeing is not a sudden formed by government crime labs. The DNA from her child, several instances in I deterioration in the quality of DNA test- work of Peter Neufeld, Barry Scheck, which samples were accidentally C

ing. It is the inevitable emergence and and their colleagues at the Cardozo Law switched or mislabeled, an instance in D

recognition of problems that existed all School Innocence Project has been which an analyst’s DNA contaminated N

along but heretofore were successfully instrumental in exposing problems, par- samples, an instance in which DNA A

hidden. In this article, I will describe these ticularly in Texas and Virginia. extracted from two different samples was underlying problems, comment on why accidentally combined into the same T they are occurring, and discuss what Surprising Frequency tube, falsely creating a mixed sample, and E defense lawyers can do about them. Of Cross-Contamination And an instance in which a suspect tested S Sample Mix-Ups twice did not match himself (probably T I

Bad Labs Another problem now emerging due to another sample-labeling error). N

One chronic problem that is now into the light is an unexpectedly high The errors documented in these ﬁles G being recognized is the uneven quality of rate of laboratory errors involving mix- are disturbing, in part, because they forensic DNA laboratories. Laboratories up and cross-contamination of DNA probably represent just the tip of an omi- vary greatly in the care with which they samples. Errors of this type appear to be nous iceberg. The documented errors validate their methods and the rigor chronic and occur even at the best DNA are, of course, those that the laboratory with which they carry them out. Quality labs. This is a problem that forensic sci- itself caught and corrected. In most control and quality assurance proce- entists have largely managed to keep instances, these errors produced unex- dures that are followed religiously in under wraps (perhaps because it is pected results that ﬂagged the problem, some labs are ignored or followed inter- always embarrassing). Practitioners have such as positive results in a control sam- mittently in others. long claimed that the rate of laboratory ple that was supposed to contain no

WWW.NACDL.ORG JANUARY/FEBRUARY 2006 11 DNA or a second DNA profile in a sam- demanded DNA tests in an ever-expand- receive scrutiny.”22 The analyst fired by ple that was supposed to be from a single ing number of cases, putting pressure on the Office of the Chief Medical person. Upon noticing such problems, labs to keep pace. Some labs have Examiner in New York used a similar labs typically throw out the results of that become high-tech sweatshops in which strategy and was caught only after con- test and start over. Accordingly, analysts analysts are under pressure to maintain fiding her activities to another analyst, usually argue that most of the incidents productivity.20 In this environment, the who then made an anonymous report to documented in these files are “not really failure of a scientific control can be a big the laboratory managers. errors” because they did not affect the problem for a DNA analyst — it forces Analyst Sarah Blair of Orchid- final results of the analysis and, in fact, the analyst to redo the entire case, Cellmark used yet another approach. She are evidence that “the system is working” putting him or her behind schedule. manipulated the computer files pro- G to detect errors when they occur. Furthermore, the presence of DNA duced by the genetic analyzer, replacing N

I However, the same processes that in an extraction blank can be embarrass- the computerized results for problematic

T cause detectable errors in some cases can ing for an analyst because contamina- control samples with the results of “clean S cause undetectable errors in others. If tion is often the result of sloppy labora- controls” from other cases. This manipu- E DNA from a suspect is accidentally tory technique. Having to redo the anal- lation was uncovered when another T

transferred into a “blank” control sam- ysis can also lead to uncomfortable Orchid-Cellmark analyst who reviewed

A ple, it is obvious that something is questions about why the analyst needed Blair’s work noticed that the same con-

N wrong; if the suspect’s DNA is acciden- two or more tries to get the test right. trol ﬁle (which happened to contain an

D tally transferred into an evidentiary sam- DNA tests themselves are viewed as unusual anomaly) appeared in two dif-

ple, the error is not obvious because infallible, so any problem that occurs in ferent cases. According to Robin Cotton,

C there is another explanation — i.e., that testing tends to be attributed (fairly or a Technical Director for the lab, a subse- I the suspect is the source of the eviden- not) to the analyst’s incompetence. quent review of computer ﬁles in Blair’s S tiary DNA. Errors that incriminate a sus- Consequently, a single mistake can end cases found approximately 25 instances N 23

E pect are unlikely to be detected as errors; an analyst’s career. After an accidental in which Blair had substituted controls.

R they are likely to be treated as incrimi- sample switch caused an embarrassing Dr. Simon Ford, an independent DNA nating evidence. Consequently, the fat false incrimination in a North Carolina consultant in San Francisco who has O

F ﬁles full of errors that a lab was able to case, for example, the lab director reviewed some of Blair’s work for the Los

catch should not be taken as reassuring reported that “the analyst working this Angeles County Public Defender’s

N evidence that “the system is working.” case was removed from casework Ofﬁce, has reported ﬁnding additional I

They are a warning signal that we need to through retirement” and “will not be instances of data manipulation in Blair’s

S worry about errors the lab did not catch. reemployed by this agency to conduct cases based on close examination of 21 24

M any type of forensic testing.” printouts of the computer data. Dr.

E Dishonest DNA Analysts So what is a DNA analyst to do if Ford also recently reported discovering a

L A third problem now emerging is problems (such as positive results in case in which an Arizona DNA analyst

B dishonest DNA analysts who falsify test blank control samples) occur too often? surreptitiously manipulated computer

O results. I suspect this third problem is For an analyst who thinks that the test ﬁles in order to cover up an error involv- 25 R closely related to the second problem: results are right anyway, it must be very ing mislabeling of samples.

P DNA analysts are faking test results to tempting just to hide the problem. This

T cover up errors arising from cross-con- can be done in a number of ways. The What Defense Lawyers tamination of DNA samples and sample DNA analysts in the Houston Police Need To Do N

E mix-ups. Crime lab came up with an easy solution Criminal defense lawyers can play a

C Given the unexpectedly high fre- — they simply failed to run extraction key role in further exposing DNA testing

E quency of contamination in DNA test- blanks (although they claimed in testi- problems and advocating for laboratory

R ing we have just discussed, it is interest- mony that they had run all necessary reform. The ﬁrst step, when handling a ing, and not at all surprising, that the controls). case involving DNA evidence, is to ﬁght G major form of fakery discovered to date According to the Inspector General relentlessly for full disclosure of the

N involves control samples known as of the Justice Department, FBI analyst underlying laboratory records and for I extraction blanks that are designed to Jacqueline Blake followed only a slightly appointment of an independent expert D detect contamination. These samples are more subtle approach. Although she to help review those records. As noted N supposed to contain no DNA. When prepared extraction blanks along with above, some of the worst laboratory A they produce positive results, it indicates other samples and recorded the creation work, such as ﬂawed DNA testing by the T

S there was a problem — DNA somehow of these samples in her notes, she Houston Police Department Crime

R ended up in a sample where it did not dumped the portion of these samples Laboratory, has occurred in jurisdictions

E belong. If that happened to a control that might have contained contaminat- where laboratory work has rarely

D sample, it could also have happened to ing DNA before sending the samples received outside scrutiny. Analysts who

N other samples, so the analyst must throw through the computer-operated genetic know that no one will ever check or

U out the whole test and start over. analyzer that typed the DNA. challenge their conclusions tend to The temptation to fake controls Interestingly, although Blake’s miscon- become sloppy, to cut corners, and to probably arises partly from production duct could have been detected by close shade their findings in ways they find pressures and partly from the collision examination of the electronic files pro- convenient. The criminal defense bar has between the public image of DNA test- duced by the genetic analyzer, no one an important role to play in maintaining ing as infallible and the reality that it is either inside or outside the Bureau the quality of forensic science, but can easier than one might expect to botch a checked this aspect of her work. Hence, only play that role if defense lawyers DNA test by cross-contaminating sam- “Blake’s record of contamination-free forcefully assert their clients’ Sixth ples. Police and prosecutors have testing for more than two years did not Amendment right to examine the evi-

12 WWW.NACDL.ORG THE CHAMPION dence against them. stances, amount to “evidence tampering” error in the case, DNA from the murder is absurd given that defense experts work victim accidentally ended up in a control Discovery Of Electronic Data only with a copy of the original data. By sample. This error was detected because A key aspect of discovery in DNA analogy, if the police disclosed digital the control sample was a “blank” which cases is the electronic data produced by photos of the crime scene, it would hard- was supposed to contain no DNA. the computer-controlled genetic analyz- ly be “data tampering” for a defense However, it was merely happenstance ers that are currently used to “type” expert to manipulate the digital images that the accidental transfer of DNA DNA samples. Analysis of the computer in an effort to enhance them or bring ended up in a blank control rather than

files can not only reveal undisclosed speciﬁc details into focus. Nor should the another sample in the same batch. If the U

problems and support alternative inter- government be able to dictate the soft- victim’s DNA had instead ended up in N pretations of the ﬁndings, but also, as ware and “analysis parameters” used by a one of the samples from the suspect’s D discussed above, these ﬁles can be crucial defense expert to examine the govern- house, I believe that the error would not E

for detecting instances of scientific ment’s digital data. The use of proper have been detected and would have led R

fraud, such as that committed by analysis parameters might well become to a false laboratory report saying that S

Jacqueline Blake and Sarah Blair. One of an issue if and when the defense decided the murder victim DNA had been found T the ironies of the Sarah Blair case is that to offer into evidence the results of its on an item collected in the suspect’s A

Orchid-Cellmark has long resisted pro- analysis of the electronic data. But to house. Defense lawyers need to think N viding full disclosure of electronic ﬁles deny access to digital data on grounds carefully about the potential for such D in its cases. Although Orchid-Cellmark that the defense might analyze it improp- errors because experience shows that I has generally, when asked, made a partial erly eviscerates the right to discovery. they can and do occur. N disclose of electronic ﬁles, their disclo- In several instances accidental G sure is less complete than that routinely Taking Contamination Seriously cross-contamination of DNA samples in R provided by other labs, including the After obtaining complete laboratory a laboratory have caused false “cold hits” E

FBI. According to Dr. Simon Ford, Sarah records in the case, defense lawyers need (i.e., false matches to an individual who C

Blair’s misconduct might well have been to review them carefully with the assis- was identiﬁed only because his or her E noticed earlier by independent analysts tance of an expert in order to identify all DNA profile was in a government N

had Orchid-Cellmark been more open possible explanation for the laboratory database). For example, the Washington T about its work. Hence, the Blair case is ﬁndings that might be consistent with State Patrol laboratory accidentally con- another instance in which absence of innocence. In light of the frequency of taminated samples from a rape case with P external scrutiny helped hide problems. contamination and mislabeling prob- DNA from the reference sample of a R The great majority of forensic DNA lems, it is particularly important to con- juvenile felon. Luckily, the case in ques- O laboratories in the United States will, on sider whether accidental cross-contami- tion was an old one. Because the juvenile B L

receiving a proper request, provide com- nation or mislabeling of samples could offender had been a young child when E

plete copies of the electronic data in a account for incriminating ﬁndings. the rape occurred he could not plausibly M case. The electronic ﬁles are typically One sure pathway to a false incrim- be connected to the case. According to S

burned onto a CD-Rom in a simple ination is accidental contamination of the lab’s Contamination/Extraneous operation similar to copying digital an evidentiary sample with DNA from a DNA Log, “it was determined that the I music or photos. Although the ﬁles are suspect’s reference sample. Given the felon’s sample was being used as a train- N

in a proprietary format that is readable known danger of cross-contamination ing sample by another analyst” when the F only by software created by vendors of among samples being processed togeth- rape case was being analyzed. In the O

the genetic analyzers, a number of inde- er as a batch, most DNA laboratories Orange County, California, Sheriff- R

pendent experts have access to this soft- take care to process the evidentiary sam- Coroner’s crime lab an analyst acciden- E ware and can use it to reanalyze the ples at a different time or place than ref- tally cross-contaminated samples from N results from the genetic analyzer, check erence samples. However, some labora- two rape cases being processed at the S the lab’s interpretations, and look for tories (the bad ones) insist on processing same time, producing another false cold I other problems. reference samples and evidentiary sam- hit. False cold hits due to accidental C

In a few jurisdictions government ples from a case all at the same time, a cross-contamination of samples from D

labs have adamantly refused to disclose practice that seems irresponsible, even different cases have also been reported in N 27

electronic DNA testing data, citing fears outrageous, given the danger that a lab- New Zealand and Australia. A

that defense counsel and their experts oratory accident could produce a false In one particularly interesting will misuse it. The Deputy Director of incrimination. Australian case, DNA on the clothing of a T the Michigan State Police issued a state- A false incrimination can also occur murdered toddler named Jaidyn Leskie E ment on May 12, 2005 opposing “the through cross-contamination among was linked, via a “cold hit,” to a young S allowance of releasing raw electronic evidentiary samples. Even labs that are “mentally challenged” woman who lived T I data for subsequent manipulations careful to test reference samples sepa- hundreds of miles away and who, by all N using software and parameters not vali- rately from evidentiary samples often accounts, had never left her own village. G dated by the Michigan State Police process all of the evidentiary samples Police could ﬁnd no way to link the Forensic Laboratory” and declaring that from a case together, creating the poten- young woman to the Leskie murder and “it is the position of the Michigan State tial for false matches. I recently reviewed at ﬁrst dismissed the “cold hit” as an Police Forensic Science Division that any a case processed by the Los Angeles “adventitious” (coincidental) match. release of this (sic) data for processing Police Department DNA laboratory in However, a coroner’s investigation estab- with non-validated parameters is tanta- which samples from a bloody murder lished that DNA from the young woman mount to evidence tampering.”26 scene were being processed in the same had been processed through the same The claim that defense reanalysis of batch as samples from items collected in laboratory at about the same time as the electronic data could, under any circum- a suspect’s house. Due to an analyst’s toddler’s clothing. The young woman

WWW.NACDL.ORG JANUARY/FEBRUARY 2006 13 had allegedly been the victim of a sexual would be well advised to investigate DNA labs (Professor Dan Krane of assault involving a condom. Although carefully whether a laboratory accident Wright State University) to the scientiﬁc laboratory personnel maintain that acci- could explain the test results. In such review panel. Fittingly, upon being dental transfer of samples between cases cases it is crucial to know where your appointed, Krane told journalists that he is impossible, it now appears almost cer- client’s DNA has been and to ascertain intended to be “an advocate of open- tain that the young woman’s DNA from whether samples taken from your client ness,” saying he would explore ways to the outside of the condom accidentally in connection with any other matters make sure laboratory work could be contaminated samples from the toddler’s might have crossed paths in the labora- independently evaluated.33 clothing. The alternative explanation — tory with the evidentiary samples that This very positive development in that there was a coincidental match were later found to contain incriminat- Virginia would never have happened G between the young woman and another ing DNA from your client. without the effective advocacy by mem- N

I person who was involved with the tod- bers of the defense bar. Steven Benjamin

T dler’s murder — has become increasing- Publicizing Problems and Betty Layne DesPortes of Richmond S ly unlikely because additional DNA test- Defense lawyers are often among the played a particularly important role in E ing, and re-analysis of the lab’s electronic ﬁrst to know about problems in forensic advocating for reform, as did Peter T

data, has reduced the likelihood of such a laboratories because they encounter these Neufeld and Barry Scheck of the 28

A coincidence to one in many trillions. problems when reviewing cases. While Innocence Project. The time is now ripe

N The facts of some recent cases in the the primary obligation, on ﬁnding evi- to continue the reform movement and

D United States have also raised suspicions dence of a lab problem, is of course to use help spread Professor Krane’s commit-

about false cold hits due to contamina- that evidence to advocate effectively for ment to scientiﬁc openness to regressive

C tion across cases. For example, in 2002, the client, it is vital that defense lawyers jurisdictions such as Michigan, where I while investigating the 1969 murder of also bring such evidence to the attention state ofﬁcials continue to pursue policies S University of Michigan law student Jane of the broader legal and scientiﬁc com- designed to cloak the work of forensic N

E Mixer, the Michigan State Police Crime munity so that underlying problems can laboratories in secrecy. Another reform

R Laboratory in Lansing found DNA of two be recognized and addressed. An excel- worth pursuing is external blind proﬁ- men on her clothing. The proﬁles were lent way to share evidence about lab ciency testing of forensic DNA laborato- O

F searched through a database and problems with other defense lawyers is to ries. Although most DNA laboratories

matched two Michigan men, Gary upload copies of relevant documents participate in periodic proﬁciency tests,

N Leiterman and John Ruelas. Police imme- (e.g., transcripts, lab reports, lab notes) to these are open tests in which the analysts I

diately suspected that Leiterman and the on-line Forensics Library operated know that they are being tested. These

S Ruelas had been involved in the murder, jointly be the NLADA and NACDL at tests have also been criticized as too easy 32

M but there was a problem — Ruelas was www.nlada.org/Defender/forensics. to detect problems that might arise in

E only four years old when Mixer was killed tough casework. A better approach is to

L and had been living with his parents in Potential Reforms occasionally ask the lab, without the

B another city. According to news accounts, Defense lawyers should also play an analysts’ knowledge, to analyze a simu-

O police could ﬁnd no link between young active role in advocating for laboratory lated case that is constructed to test the 29 R Ruelas and Mixer. That did not deter reform. Perhaps the sole positive aspect lab’s performance.

P Washtenaw County Assistant Prosecutor of the recent spate of DNA testing prob- In 1992, the National Research

T Steven Hiller who charged Leiterman lems is that legislators, judges, and even Council called for external, blind DNA with the murder. Hiller “created a sce- prosecutors are gradually becoming proﬁciency tests “that are truly represen- N

E nario placing a young Ruelas at the [mur- aware of underlying problems in foren- tative of case materials (with respect to

C der] scene as a chronic nose-bleeder sic science, making this an opportune sample quality, accompanying descrip- 30 34 E whose blood dropped on Mixer.” There time to press for reforms. tion, etc.).” Thereafter, the Federal DNA

R is, however, another possible explanation One such reform is the creation of Identiﬁcation Act of 1994 required the for this “cold hit.” Examination of labora- independent commissions to supervise director of the National Institute of G tory records revealed that known samples the operation of forensic laboratories. Justice (NIJ) to report to Congress on the

N of DNA from both Leiterman and Ruelas New York state’s successful forensic sci- feasibility of establishing such a testing I were being processed in the Michigan ence commission is a good model. It is program for DNA laboratories. The D State lab on the same day as the old sam- encouraging that, in the wake of crime National Institute of Justice funded a N ples from the Mixer murder.31 Both men lab scandals in those states, Texas and major study of this issue in which small- A were being tested in connection with Virginia have both recently adopted sim- scale blind proﬁciency tests were con- T

S other cases unrelated to the Mixer mur- ilar legislation. The Virginia legislation ducted to assess their practicality and

R der. Although the Michigan State labora- creates a scientiﬁc review panel with the costs. The study found that blind proﬁ-

E tory maintains that cross-contamination authority to review laboratory opera- ciency testing is possible, although some- 35 D of samples across cases was impossible, it tions, adopt qualiﬁcation standards for what difﬁcult to administer. The esti-

N seems a very strange and unlikely coinci- the lab director and other staffers, and mated annual cost of administering two

U dence that two men who, according to the establish an audit process to be used blind proﬁciency tests (involving simu- prosecutor, were present when Mixer was when errors occur. The panel will also be lated cases) to each of the 150 DNA test- murdered in 1969 just happened to have available to review lab reports and test ing laboratories in the United States was their DNA tested (for other cases) on the results at the request of the governor or only $450,000 to $3,020,000 which seems very same day as samples from the Mixer lab officials. Virginia Governor Mark easily affordable. The move toward exter- case were tested. Leiterman was neverthe- Warner (D) signaled his seriousness nal blind proficiency testing lost less convicted of Mixer’s murder in 2005. about improving the quality of forensic momentum, however, when an NIJ sci- Lawyers who represent clients who DNA testing in the state by appointing a entiﬁc advisory board concluded that are incriminated through “cold hits” prominent academic critic of forensic such a program would not be worth the

14 WWW.NACDL.ORG THE CHAMPION trouble.36 Ironically, the key arguments http://seattlepi.nwsource.com/specials/crim Shaler, former Director of the OCME DNA advanced for this position were that lab- elab/ Laboratory. oratory errors occur so rarely that profi- 4. See stories collected at: http://crime- 17. Associated Press, Worker in Army lab ciency tests would be unlikely ever to andscience.journalnow.com; see also, may have falsified DNA test result.Aug.27, catch them and that effective alternative Phoebe Zerwick, State crime lab is faulted: 2005; Memorandum For All Staff Judge methods exist for maintaining the quali- Lawyers’group calls for probe,cites DNA errors Advocates Re: Brady Notice by Lisa Kreeger, ty of DNA laboratory work. Both of in three cases, WINSTON-SALEM JOURNAL, July 20, Staff Attorney, Department of the Army, these arguments have been cast in seri- 2005. October 17, 2005 (on file with the author).

ous doubt by subsequent events. In light 5. See Frank Green, Mistakes by state 18. See, Michael Bromwich, Third report U

of the serious problems that have recent- DNA ﬁrm alleged: The Illinois State Police,‘out- of the independent investigator for the N ly come to light in forensic DNA labora- raged’ by ﬁndings, end their contract with the Houston Police Department Crime Laboratory D tories, it is high time to revisit the issue of firm.RICHMOND TIMES-DISPATCH, August 20, and Property Room. Available online at: E

external blind proﬁciency testing. 2005. www.hpdlabinvestigation.org/. R

6. Tom Jackman, Paternity Suit Raises 19. DNA Advisory Board Quality S

Conclusions Doubts About DNA Tests: Va. Judge Rejects Assurance Standards for Forensic DNA T DNA evidence is difﬁcult to chal- Results, Questions Lab Work in Case of D.C. Testing Laboratories, Standard 14.1.1. A

lenge in the courtroom because most Hair Salon Owner.WASHINGTON POST, August Available online at: www.cstl.nist.gov/ N

people think it is virtually infallible. It is 21, 2005; C01. div831/strbase/dabqas.htm. D

not just jurors, fed on a media diet of 7. See, David Chanen, Defense attorneys 20. In a recent case in Washington, D.C. I CSI-style fantasies, who think so. Most raise concerns about DNA sample mix-up. involving an allegedly inaccurate DNA pater- N members of the academic and legal MINNEAPOLIS STAR-TRIBUNE, May 20, 2005; BCA nity test, evidence showed that LabCorp, G community believe it as well. Even crime lab under the microscope, KSTP-TV which conducts over 100,000 paternity tests R scholars who are critical of other areas of NEWS (at www.kstp.com/article/stories a year, “has only ﬁve people reviewing the E

forensic identification science have /S8367.html?cat=1). data and making paternity determinations C

argued that DNA is an exception—call- 8. Letter from Jerry Richardson, Crime — with one supervisor testifying that he E

ing DNA testing “a model for scientiﬁ- Laboratory Director, NC State Bureau of issues an average of one paternity report N 37 cally sound identiﬁcation science.” Investigation to Ralph Keaton, Director of every four minutes during a 10-hour shift.” T

While there is no doubt that DNA ASCLD/LAB, May 23, 2005 (“During the Tom Jackman, Paternity Suit Raises Doubts testing rests on a stronger scientiﬁc foun- course of the analysis the analyst extracted About DNA Tests: Va. Judge Rejects Results, P dation than many other forensic disci- DNA standards from the suspect and victim Questions Lab Work in Case of D.C. Hair Salon R O plines,38 recent events have proven that and switched the samples…which led the Owner. WASHINGTON POST, Sunday, August 21, DNA evidence is hardly infallible. The analyst to conclude incorrectly that the sus- 2005; C01 B pect was the source of blood found at the 21. Letter from Jerry Richardson, supra L solid scientiﬁc foundation for DNA test- E

ing is no guarantee that DNA tests will be crime scene”)(letter on ﬁle with the author). note 8. M carried out in a reliable manner that pro- 9. William C. Thompson, et al. How the 22. U.S. Department of Justice, Ofﬁce of S

duces accurate results. Bad laboratory probability of a false positive affects the value the Inspector General, The FBI DNA

work is all too common and laboratory of forensic DNA evidence. 48 J.FORENSIC SCI. 47 Laboratory: Review of Protocol and Practice I accidents and errors can occur even in (2003)(discussing false incrimination of rape Vulnerabilities, May 2004. Available on-line N

good labs. Whether DNA evidence is suspects in Philadelphia, San Diego, and at: http://www.usdoj.gov/oig/special/0405 F

trustworthy is a question that must be Tulsa). /ﬁnal.pdf. O

examined carefully in each case. And that 10. Glen Puit, DNA evidence: Ofﬁcials 23. Letter from Robin Cotton, Director, R

challenging task falls ultimately on the admit error, dismiss case (LV lab put wrong Technical Forensic Science, Orchid-Cellmark E

shoulders of lawyers who represent name on sample). LAS VEGAS REVIEW JOURNAL, to Ralph Keaton, Executive Director, N

clients incriminated by DNA tests. April 18, 2002. ASCLD/LAB,dated Sept 20,2004 (on ﬁle with S

11. See text preceeding note 27, infra. the author). I Notes 12. Roma Khanna, Retesting of crime lab 24. Simon Ford, Fraud detection C

1. The HOUSTON CHRONICLE maintains an work in question.HOUSTON CHRONICLE, Dec 6, through case reviews. Presentation at D

archive of articles about the scandalously 2004. Forensic Bioinformatics 4th Annual N

bad work of the Houston Police Department 13. A “cold hit” occurs when a DNA pro- Conference: The Science of DNA Proﬁling: A A Crime Laboratory at www.chron.com/con- ﬁle found in an evidentiary sample in a case National Expert Forum, Dayton, Ohio, Aug.

tent/chronicle/special/03/crimelab/index.ht with no obvious suspects is found to match 13, 2005 (Available on-line at: T

ml. a DNA proﬁle in a government database, http://www.bioforensics.com/confer- E 2. The VIRGINIAN-PILOT and the RICHMOND such as a database of convicted offenders or ence05/FBS_Dayton_2005_Fraud.pdf) S T TIMES-DISPATCH have published a series of a database of samples from other unsolved 25. Id. I

news article and editorials about DNA test- crimes. 26. Letter from Lt. Colonel Timothy J. N

ing problem in the Virginia State Division of 14. See, U.S. Department of Justice, Yungfer, Deputy Director, State Services G Forensic Sciences. See, e.g., Confusion over Office of the Inspector General,The FBI DNA Bureau, Michigan State Police to Stuart DNA a threat to Justice,VIRGINIAN-PILOT,Aug. Laboratory: A Review of Protocol and Dunnings, President, Prosecuting Attorneys 29, 2005; Frank Green, Study will assess Practice Vulnerabilities, May 2004. Available Association of Michigan, May 12, 2005 (on whether errors in Washington case are online at: http://www.usdoj.gov/oig/spe- file with the author). 'endemic to the system.' RICHMOND TIMES- cial/0405/index.htm 27. For an account of the false cold hit, DISPATCH, June 14, 2005; Alarming indifference 15. Laura Cadiz, Md.-based DNA lab fires and other DNA testing errors in New from crime lab boss, VIRGINIAN-PILOT, May 10, analyst over falsified tests, BALTIMORE SUN, Nov. Zealand,see Michael Strutt,Legally scientific? 2005. 18, 2004. A brief history of DNA evidence in the crimi- 3. See news stories collected at: 16. Author’s interview with Robert nal justice system. June 9, 2001 (available on-

WWW.NACDL.ORG JANUARY/FEBRUARY 2006 15 line at: http://www.justiceaction.org.au Lab Y”).If you have only a hard copy of a doc- Koehler,The coming paradigm shift in foren- /actNow/Campaigns/DNA/pdf_files/02_Leg ument,and are not able to scan it,you can fax sic identification science. 309 SCIENCE 892 al.pdf) it to (202) 824-2929. Once uploaded, docu- (Aug 5, 2005). 28. See,William C.Thompson, Report to ments are reviewed by the staff of the 38. There is a solid scientific basis for the Victoria State Coroner Regarding the Forensics Library, and then posted. There are characterizing and estimating the frequency Death of Jaidyn Leskie, December 3, 2003; three possible levels of access to documents of matching DNA profiles that simply does William C. Thompson, Supplemental Report in the Library.Documents can be made avail- not exist for matching fingerprints, to the Victoria State Coroner, Jan 29, 2004; able to anyone who comes to the website, or bitemarks, toolmarks, etc. ■ Dan Krane, Report: Victoria State Coroner’s available only to those who attest that they Inquest into the Death of Jaidyn Leskie, are defense lawyers, or available only on a About the Author G December 4, 2003. These reports are avail- need to know basis to people approved by N

I able online at: www.scientiﬁc.org. the Library staff. William C.Thompson, J.D., Ph.D., is a pro- T 29. Maryanne George, Murder case mys- 33.Christina Nuckols,Governor appoints fessor in the Depart- S tery deepens. DETROIT FREE PRESS, Jan 15, 2005. panel to oversee Va’s.crime lab. VIRGINIAN-PILOT, ment of Criminolo- E 30. According to news accounts, Hiller August 9, 2005.

T gy, Law & Society at

offered no evidence to support this theory. 34. National Research Council, DNA the University of

A Liz Cobbs, Judge raises possibility evidence Technology in Forensic Science (1992). California, Irvine

N may have been contaminated at State Police 35. Joseph L. Peterson, George Lim, and a member of ,ANN ARBOR NEWS, May 11, 2005. Monica Ho, Yingyu Chen & R.E. Gaensslen, D lab the California bar. 31. Author’s interview with Professor The Feasibility of External Blind DNA He is co-chair of C Dan Krane (a defense expert in the case). Proﬁciency Testing I,Background and Findings.

I NACDL’s Forensic Also,Thersa Mask,Mixer’s dad is clear on one 48 J.FORENSIC SCI. 1 (2003); Joseph L. Peterson, S Evidence Committee. thing, DETROIT FREE PRESS, July 13, 2005. George Lim, Monica Ho, Yingyu Chen & R.E. N 32.Documents in any digital format can Gaensslen, The Feasibility of External Blind E William C. Thompson

R be uploaded to the library from any comput- DNA Proﬁciency Testing II, Experience with Department of Criminology, er with access to the web. As you upload Actual Blind Tests. 48 J.FORENSIC SCI. 1 (2003). O Law & Society

F each document, you can ﬁll out a convenient 36. The author of this article was a University of California digital form explaining what the document is member of that advisory panel, which con- Irvine, CA 92697 N and why it is important (e.g., “This is a tran- sisted primarily of forensic scientists, and

I 949-824-6156

script in which analyst from Lab X admits to cast the sole vote in favor of recommending Fax 949-824-3001 S accidentally switching samples during DNA a program of external blind proficiency tests. E-MAIL [email protected]

M testing”or “This is a corrective action ﬁle from 37. Michael J. Saks & Jonathan J. E L

B Show Your NACDL Membership With Pride in Court — and at Home! O R P

Announcing the NACDL Wristband inscribed with: T N E “LIBERTY’S LAST CHAMPION” C E

R • $10.00 each

G • Orders of 10 or more: $7.50 each N I Buy one — buy 10, and give them to your colleagues D N A T S R E D N U

All proceeds support NACDL’s Foundation for Criminal Justice To order: Visit www.nacdl.org/shop

16 WWW.NACDL.ORG THE CHAMPION