Forensic Uncertainty

The forensic uncertainty principle may seem complicated. It is really simple when you think about it. In every measurement, there are sources of error that contribute to uncertainty in the reporting of that measurement. The error bars that contain the true value or concentration of a chemical substance that one wants to identify and measure (defined as an "analyte" in the vocabulary of analytical chemistry) have to be established based on collecting, studying, analyzing and calculating data using rigorous adherence to the scientific method and good statistical sampling and calculation. In a prosecution for drunk driving, for example, the prosecutor can enter chemical test evidence of your alcohol content in order to convict you. (MCL 257.625a(6) (b)ii.) That evidence is not reliable because it is analyzed and reported as a single number such as .29 or .08, by the Michigan State Police Toxicology lab. The fundamental aspect of metrology or the science of measurement is that in every measurement there contains a range of values, any one of which may include the true value of the thing being measured. That range of values is called the uncertainty budget. The lack of uncertainty budgets in forensic measurement is among the critiques of the National Academy of Sciences in its critical, eye-opening report "Strengthening Forensic Science: A Path Forward." This report was commissioned by the United States Congress in response to the unreliability of forensic evidence in the nation's courtrooms. Despite the ray of sunlight that it beamed onto the American criminal justice system, the NAS report has received spotty recognition and spotty follow-up in my view. The NAS report can be found at http:// www.nap.edu. The NAS report's findings are also part of the ISO 17025 requirements for building uncertainty budgets and reporting results within statistically established error bars. The body that accredits crime labs across the country, the American Society of Crime Lab Directors-LAB committee (ASCLAD-LAB) pronounced that labs now are required to employ measurement uncertainty in estimating concentrations of analytes in subject samples. The uncertainty measurement should be included for both quantitative (how much is there) and qualitative (is it there) analysis. Much like parents or loved ones who enable addicts and abusers to continue using their drug of choice, we are only enabling wrongful convictions and junk science when we KNOW what the problems are yet fail to change our behaviors. Sometimes you have to be brave and conduct an intervention for a friend or loved one who is an abuser. Sometimes you have to be brave if you are afflicted with addiction and you are called out on your illness and you fight the difficult fight, inch-by-inch and day-by-day, to live a sober lifestyle. We in the criminal justice system are the abusers, the NAS report is our intervention. That brings me to a case in Mason County Michigan in which 79th Judicial District Judge Peter J. Wadel excluded a blood test for lack of reliability under MRE 702. Statistically, the measurement had no meaning because the lab witnesses insisted that it represented a single or "absolute" value. It predicted nothing and was simply a statement from the prosecutor and crime lab. The judge's opinion and order can be found at http://www. nicholslawyers.com Judge Wadel spent over 2 years analyzing whether or not the Michigan State Police Crime Lab could validly demonstrate controlled statistical analysis of its data and reliable application of that data to an uncertainty budget. Time and time again during the hearings, the lab director insisted that there was no uncertainty budget and that there will not be one unless the lab is forced to come up with one. The judge's opinion was straightforward. It did not contain any extraordinary new legal theory or scientific principle. Judge Wadel said that the Lab would not produce an uncertainty measurement. Without the uncertainty budget, the evidence was misleading because if something is measured 100 times there will likely be 100 different results from a statistical standpoint. We have the tools in the Michigan criminal justice system. Fortunately, the Michigan Supreme Court amended the rules of scientific or "expert" testimony in 2003 when it promulgated a new version of MRE 702. Testimony that is expert or scientific in nature is inadmissible unless it is both relevant and reliable as defined by the rule. Because the Supreme Court's rule-making power trumps the statute on a matter of substantive legal procedure, MRE 702 trumps MCL 257.625a(6)(b)(ii). All we need to do is start looking to science to guide us in deciding whether a forensic measurement is sufficiently reliable under the principles of Daubert. Here is the alternative: we stain people for life with a conviction because it is easier than measuring data and fully informing juries. "It's just a DUI." Eventually, the whole system will come tumbling down. People will stop abiding by judgments, verdicts and probation orders because no one has faith in a system that ignores science in favor of what is easier for government-paid lab technicians who work for the state police. If we hide behind "come on, you were presumed innocent" and "this is the best imperfect system in the world." Is it? I suggest to you that our arrogance could be our downfall. Mike Nichols is author of the Michigan OWI Handbook, an adjunct professor of DUI Law and Practice at Thomas M. Cooley Law School, author of a chapter in a national text on forensic metrology. He can be reached at mnichols@ nicholslawyers.net. Published: Mon, Jun 13, 2011