NIJ Study of Near Misses and Wrongful Convictions

The National Institute of Justice recently released their study Predicting Erroneous Convictions which examines why some innocent people are wrongfully convicted while others are acquitted, using a case comparison method rather than a traditional “case study” method. Researchers at American University who compared a group of 260 cases that occurred between 1980 and 2012 where an innocent defendant was exonerated only after conviction with a control group of 200 cases where an innocent defendant was acquitted or had charges dismissed before trial.

Unlike prior studies which have focused on what causes an innocent defendant to be charged, this study’s results can be used to determine what is preventing the innocent defendant from being acquitted or having charges dismissed once in the system. This is a helpful distinction, as it allows a greater focus on criminal defense, forensic practices, and prosecutorial discretion.

A Mixed Method Approach

The study’s quantitative analysis yielded 10 primary factors that led to a wrongful conviction rather than the “near miss” dismissals and acquittals:

• A younger defendant
• A criminal history
• A weak prosecution case
• Prosecution withheld evidence
• Lying by a non-eyewitness
• Unintentional witness misidentification
• Misinterpreting forensic evidence at trial
• A weak defense
• Defendant offered a family witness
• A “punitive” state culture

A qualitative analysis of the data followed in which a panel of criminal justice experts analyzed 39 cases to determine how the ten factors were connected and whether “tunnel vision” (when law enforcement professionals focus more on building a case against one suspect at the expense of ignoring contradictory evidence) played a role.

The final report offers recommendations to prevent future wrongful convictions, including recommendations on defense practice, exculpatory evidence, eyewitness identification, false confessions, forensic error, police misconduct, weak prosecution evidence, systemic failures and tunnel vision. The report emphasizes that the interactions of these factors as much as the individual factors themselves are to blame for systemic breakdowns leading to erroneous convictions. This type of interaction requires a comprehensive approach to reform in order to prevent future errors.

Traditional Wrongful Conviction Factors

Factors that have traditionally been suggested as sources of erroneous convictions, such as false confessions,  witness misidentification, and racial bias, appeared at statistically similar rates in both sets of cases. Thus, these factors likely only increase the chance that an innocent suspect will be charged but not the likelihood that the charge will result in conviction rather than dismissal or acquittal.

Forensic Error as a Primary Factor in Wrongful Convictions

As one would expect, this study revealed that errors in forensic evidence were correlated with an increased likelihood of erroneous conviction. However, this study’s results suggest a new perspective on how forensic error impacts case outcomes. Errors in forensic testimony, rather than errors in the actual testing were the most common source of forensic error.  Forensic testimony errors include: not providing the jury with key forensic information, overstating the inculpatory nature of evidence, use of inaccurate statistics, and misstating the certainty of results and forensic techniques. The implication is that an emphasis on quality control at the interpretation and testimony stages should be added to previous policy recommendations that have focused on improving quality of forensic lab procedures.

While DNA is a widely respected as a forensic tool, this study found that the prohibitive cost of DNA testing often renders it a tool of last resort, with other less accurate forensic tools like fingerprinting, hair comparison analysis and serology testing being used first.

Recommendations for Improving Forensic Error

• Timeliness: Forensic investigation, especially DNA testing, should be conducted early in a case to rule suspects in or out, rather than later once an investigation gains steam and there is a danger of it being used more to confirm or check what police/prosecutors already believe.
• Training/Education: Prosecutors and defense attorneys need more education about forensic testing techniques and should get clarifications about results if they do not fully understand a report. Police officers in more jurisdictions need to be trained in crime scene investigation and should then be present when technicians are collecting evidence to help direct and expand the search.
• Supervision: Forensic labs should be more closely supervised and have procedures that are peer reviewed. There should be discussions of “near miss” cases, which could follow, for example, in the vein of the morbidity and mortality conferences that the medical community holds to discuss mistakes in patient care.

For more information, watch the forthcoming recording of the Wrongful Convictions seminar given by the report’s authors, found here.

BBC Knowledge Explainer DNA

Here is a 3-minute video that explains the basics of DNA’s form and function. It’s part of the BBC’s online Explainer documentary series. Though the focus of this short animated film is not forensic DNA analysis, it gives important information that attorneys or jurors need to understand before attempting to understand forensic DNA analysis. Attorneys should watch the video for their own information and also to think about whether this information is being communicated effectively to jurors by DNA experts.

Thousands of cases compromised due to faulty forensic analysis

In recent months, faulty forensic analysis has been exposed in several labs across the country. The failure of a handful of lab analysts to correctly perform forensic analysis has compromised thousands of cases. In each situation the failures are different, but they expose a lack of oversight of analyst performance in the affected labs. The following are several of the most serious failures:

Annie Dookhan

A Massachusetts chemist was accused of faking test results at the state lab and tampering with drug evidence while she tested suspected controlled substances in criminal cases. Authorities declared that Dookhan tested more than 60,000 samples involving 34,000 defendants during her nine years at the Department of Public Health lab. Over 200 convicted defendants have been released from custody while their cases are being reviewed due to Dookhan’s involvement, according to this article. One of the red flags that lead to Dookhan’s misconduct being detected was the fact that she was highly efficient at her job; she was handling an astounding number of samples compared to an average chemist. Investigators allege that Dookhan was able to accelerate her work by “dry labbing” or reporting results for analyses that she did not actually perform. Dookhan has been indicted on 27 charges, including 17 counts of obstruction of justice, eight counts of tampering with evidence, perjury and falsely testifying that she held a degree from a college or university.

Sonja Farak

Another Massachusetts chemist that worked in the state crime lab in Amherst was arrested in January and charged with evidence tampering and possession of controlled substances from the lab, according to this article. The Boston Globe reported that Farak was discovered when her supervisors were making a routine check of tested substances and found that certain substances tested by Farak had been replaced with counterfeit substances. Attorney General Martha Coakley said that both Farak and Dookhan had begun their career at the Hinton lab in Jamaica Plain. Unlike in Dookhan’s case, supervisors noticed that Farak had had a drop in her productivity. Authorities have stated said that Farak’s misconduct was quickly detected by her supervisors, limiting the scope of its impact.

Jonathan Salvador

A forensic scientist who worked as a controlled substances analyst at the Texas Department of Public Safety (DPS) was suspended when it was discovered that he issued a fraudulent report about a batch of pills, according to this post in the Grits for Breakfast blog. The report was issued without testing the pills and instead, substituted data from another sample. The DPS characterized Salvador’s work as deficient prior to this incident. Case supervisors were aware of Salvador’s poor performance and knew that he appeared not to understand the science of the work he was assigned. However, his performance was tolerated and he would often volunteer for unwanted tasks in the lab. In its internal investigation, the DPS found several additional cases where Salvador misreported results. According to Grits for Breakfast “hundreds of convicted defendants may end up having their cases overturned, either freeing them from prison or ending their probation terms.” It was reported that hundreds of samples tested by Salvador during his six years at the DPS were being retested.

Iowa analyst fired over mishandling of fingerprint evidence

A state police crime lab analyst in Iowa was fired in January due to errors in reports related to fingerprint analysis, according to this article. The lab reviewed the analyst’s 2012 cases and found at least nine cases contained errors where the analyst had incorrectly classified fingerprint evidence as unusable. The analyst’s errors were discovered in a routine internal review of cases. The investigation of the analyst’s casework continues. The analyst had been employed by the lab for sixteen years.

Mishandling of DNA Evidence in Rape Cases

The New York City Medical Examiner’s Office is reviewing over 800 cases worked by a lab technician who resigned in 2011. According to this New York Times article, reviewers have found so far that the technician failed to detect biological evidence in 26 cases when in fact existed. Additionally, in 55 cases, the lab technician failed to upload evidence from crime scenes into the state’s DNA database. The mishandling of DNA evidence led sex-crime investigators to not have available evidence that was could have been used to develop cases against rape suspects.  Supervisors also discovered sixteen pieces of evidence that had been placed in the wrong rape kits. The majority of the misplaced items were swabs sealed in paper envelopes.  This mixing of items from different cases raises concerns about cross-contamination and whether other lab protocols were ignored.

Additional cases of lab analyst misconduct are detailed in this NACDL News Release.

In sum, it is important for the attorneys to be aware of the risks of not reviewing the lab reports, including the underlying data, in all of their cases. Although the majority of labs endeavor to monitor the work of individual analysts through case reviews, the cases above indicate that supervision cannot completely deter deficient performance by individual analysts.

Developing Analyses of Biological Evidence: Predicting Eye Color, Determining the Source of Bodily Fluids, and Locating Trace Evidence Within Guns

Several studies reported in the January 2013 Issue of Forensic Science International: Genetics are of interest for future developments in forensic evidence.

Research is being done on predicting eye color, hair color, and skin color from DNA. Currently, 37 gene sequences (SNPs or single nucleotide polymorphisms) have been identified as playing a role in these traits, and, of these, six have previously been used for predicting eye color. The current study tested additional gene sequences, beyond the previously used six SNPs, in order to test the ability to keep the cost of the test in balance with any gains in predictive value and reliability. It is important to note that these tests assign likelihoods of particular eye colors of the source person and do not provide an absolute determination of that eye color. Such tests may be used in the future to help develop possible suspect profiles.

Another type of genetic material, mRNA (messenger RNA, the “template” copy of DNA coding used for assembling specific proteins within a cell), has been found useful in determining the source of biological fluids. A study of mRNA markers identified specific markers for blood, saliva, semen, and vaginal secretions. The ability to identify the tissue source of a DNA profile would have important forensic uses because currently it is not possible to determine whether the DNA profile developed from a particular item of evidence comes from, for example, a blood droplet or from skin cells previously left on the surface of that item of evidence.

Finally, a study examined the firearms used in 20 cases of homicide or suicide in which there was a close-contact shot fired. The inside of the barrels of the guns were sterilely swabbed from the front of the weapon (making sure to avoid contact with the muzzle) in each case and also swabbed from the rear of the weapon in 16 of the cases. Usable biologic material for DNA analysis was found in 17 of the guns. Furthermore, after the initial samples were collected, each gun was fired one additional time, and usable genetic material was still found in 14 of those guns. Not only may testing of genetic material from within gun barrels be used in crime scene investigation in the future, but parameters may be developed for specific gun types such that, if biological evidence is found within the barrel of a gun, investigators will be able to determine if the firearm likely was fired from within a certain distance from the victim’s body.

History of DNA Analysis Timeline – Available Online

A DNA analysis timeline is now available on the IDS Forensic Resources website.  The timeline traces the development of forensic DNA analysis and its use by the NC State Crime Lab.  It seeks to identify what technologies were available during a particular time period.  It is a helpful tool for attorneys working on older cases or post conviction cases who need to know what techniques were being used during the time period when evidence in their case was tested.

You can access the timeline here.

The ENCODE project, the Fourth Amendment, and Haskill v. Harris

The Ninth Circuit Court of Appeals, sitting en banc, heard oral arguments on whether DNA fingerprinting violates the Fourth Amendment’s protection against unreasonable search and seizure on September 19.

Plaintiffs-Appellants in Haskell v. Harris argue that collection of DNA at arrest violates their privacy interests because DNA contains not only CODIS markers that help in forensic identification by distinguishing one individual from the other, but also contain additional genetic information which plays a role in cell behavior.  This revelation is supported by the ENCODE Project, a nine-year, world-wide, federally-sponsored project which has determined that the once thought “junk” DNA actually contains intimate information such as an individual’s susceptibility to disease and physical traits such as height.  For this reason, opponents of DNA fingerprinting argue that the forensic use of DNA implicates a cognizable privacy interest because, while a DNA profile contains identifying information, it also contains much more intimate information that is sensitive and private to the individual.

You can read more here. A video of the arguments is posted here.

“Forensics on Trial” program on NOVA

NOVA (PBS) aired a program called “Forensics on Trial” on October 17, 2012, examining the crisis facing crime labs in the U.S which lack central oversight, employ few scientific standards, and have poor regulation of examiners. The program investigates how the use of fingerprint, bite mark, ballistics, toolmark, and hair analysis evidence has led to wrongful convictions. The program will examine the 2004 Madrid bombing case in which a so-called fingerprint match was used to arrest an innocent man. The fingerprint was later matched to another suspect.  The program will also discuss how mishandling of forensic evidence impacts the criminal justice system, from wrongful convictions to the O.J. Simpson trial.  You can read about the details of the program here.

More information is also available on the NOVA website where the program can be viewed online.

The “Birthday Problem”

Defense attorneys interested in learning more about DNA statistics might find the following articles interesting because the “birthday problem” is analogous to looking for partial matches in a DNA database.  The birthday problem is a classic puzzle that asks  if you had a room full of people, how many people would you need in the room to make the odds of two people having the same birthday at least 50-50. Surprisingly, the answer is just 23.

In the “birthday problem,” when you do not designate a specific birthdate, you open up the opportunity for many pairs to be made from a much smaller sampling of people than if you specified a specific birthday.  The same applies in with DNA when searching a DNA database for any set of matching loci, rather than looking to match a specific set of loci to another distinct set of loci.

This NY Times post is part a 6-part series entitled “Me, Myself and Math” and does a great job of explaining the birthday problem. For folks wanting more in-depth information about how the birthday problem is useful for understanding trawling of DNA databases for partial matches, take a look at this article by David H. Kaye.

Free at Last! Willie Grimes Exonerated by 3-Judge Panel

A 3-judge panel found Willie Grimes innocent after thirty-minutes of deliberation on Friday.  Grimes had been sentenced to life in prison in 1987 after being convicted of two counts of first degree rape and one count of second degree kidnapping.  He was paroled in May of this year after serving twenty-four years in prison.

Grimes’s case was heard in April by the Innocence Inquiry Commission, which unanimously agreed that enough evidence existed to refer his case to the 3-judge panel.  The rape kit and hairs found at the scene had been lost and were unable to be utilized in exonerating Grimes.  However, a fingerprint on a banana, which the victim had claimed her attacker may have touched, was analyzed and confirmed as belonging to someone other than Grimes. Grimes was represented by attorneys Chris Mumma, director of the North Carolina Center on Actual Innocence, and Robert Campbell of Hickory.

You can read more here, here, or here.

NIJ Has Made Available New Forensic Technical Reports

The National Institute of Justice has published several reports on novel techniques that are being investigated in order to improve forensic analysis. Take a look at the reports below to learn about some of the latest techniques that are being developed and to get a forecast of what techniques you may see coming soon to a forensic lab near you!

Application of Machine Learning to Toolmarks: Statistically Based Methods for Impression Pattern Comparisons (pdf, 99 pages)

Researchers created a database of 3D striation and impression patterns on Glock fired cartridge cases, screwdriver and chisel striation patterns. They attempted to objectively associate the toolmarks with the tools that created them using principal component analysis, canonical variate analysis, and support vector machine methodology. Researchers were able to estimate an error rate for toolmark identification using these techniques and a confidence level was assigned. Researchers suggest that this methodology is a useful means of gauging the quality of a toolmark “match.”

Application of Raman Spectroscopy for an Easy-to-Use, on-Field, Rapid, Nondestructive, Confirmatory Identification of Body Fluids (pdf, 80 pages)

Research indicates that new nondestructive, confirmatory testing could identify bodily fluids, including dried fluids, with a near 100% accuracy rate using a Raman microscope equipped with advanced statistics for rapid mapping of pure bodily fluids. The research further proposes that mixed samples can be identified if the samples are not completely mixed. This nondestructive testing, if adopted in the future, could help in the preservation of crucial DNA evidence. It also offers a heightened level of confidence in the results obtained due to its confirmatory nature and level of accuracy.

Establishing the Quantitative Basis for Sufficiency Thresholds and Metrics for Friction Ridge Pattern Detail and the Foundation for a Standard (pdf, 53 pages)

Motivated by both the Daubert ruling and the findings of the National Academy of Sciences, Strengthening Forensic Science in the United States: a Path Forward (2009), researchers from multiple disciplines collaborated to develop a scientific foundation for fingerprint image quality, with a focus on latent prints.  Researchers followed an experimental approach and then performed statistical validation of their results. Their research yielded several noteworthy results, most salient being detection of unique features that can assist fingerprint examiners in drawing statistical likelihoods of a given feature. This project bolsters claims that existing methods of analysis are more of an art than a science and therefore more susceptible to human error.

Filling a Critical Need by Establishing a Fully Functioning, CODIS Dedicated Laboratory (pdf, 101 pages)

The state of Wyoming requires maintenance of a database of offender samples which historically have been processed via outsourcing to private agencies.  Through funding from both the NIJ and Wyoming State legislature, Wyoming has established their own CODIS lab to process offender samples for entry into CODIS. Currently, samples processed through the lab pass through to CODIS at a 95% success rate. Through creation of the state-run lab, sample processing time has been cut from more than two years to less than sixty days.

Implementation of a DNA Triage and Analysis System Dedicated to Increasing the Throughput of High Volume Crimes in a Forensic Laboratory (pdf, 133 pages)

Through the efforts of both the Orange County California crime lab and the Orange County District Attorney’s office, the crime lab implemented a team-oriented approach to processing DNA from property crimes. The approach employed a highly automated DNA processing method that the crime lab anticipated would free up time from the processing of property crimes, which are extremely high in volume, to devote to dealing with violent crimes. As anticipated, the new handling method decreased the turn around time on both property and violent crimes.

Taq Mutants Engineered for Forensics (pdf, 44 pages)

Researchers used novel genetically-engineered enzymes and a new protocol for DNA analysis to attempt to reduce false negative results and improve the efficiency of DNA testing. These novel enzymes and new protocol would work better with forensic samples that contain residual blood, soil, or other substances that inhibit forensic DNA analysis. Researchers found that their technique outperformed the techniques currently used in forensic DNA analysis.

Use of Scanning Electron Microscopy/Energy Dispersive Spectroscopy (SEM/EDS) Methods for the Analysis of Small Particles Adhering to Carpet Fiber Surfaces as a Means to Test Associations of Trace Evidence in a Way that is Independent of Manufactured Characteristics (pdf, 77 pages)

Typically, very small particles (VSP) are ignored by forensic science. However, researchers have made the first steps towards developing a method for collecting and analyzing VSP as part of trace evidence. Researchers have successfully collected VSP from carpet fibers and demonstrated the difference between carpet fibers themselves and the VSP adhering to them. Researchers urge that VSP can prove useful in the area of trace evidence by comparing VSP from crime scenes or suspects to any items of physical evidence.