Pt 2: Imprecise DWI Tests: Good Enough for Government Work?

Posted On September 13, 2011 by Daniel Koewler

Last week we told you that the scientific community requires blood, breath and urine tests be reported not as a certainty, but with a confidence interval that has a high probability of containing the true alcohol level. (This is not to be confused with results involving mistakes or blunders; uncertainty assumes that no mistakes or blunders have been made.) Therefore, due to the uncertainty in every measurement, the state must report a DWI alcohol test results with a range, not as a single number.

This raises some interesting questions for drivers in Minnesota who face criminal penalties or civil license revocation, or vehicle forfeiture.

Are there established confidence intervals for Minnesota DWI tests?

Neither the Minnesota Commissioner of Public Safety nor the Minnesota Bureau of Criminal Apprehension (BCA) has published established confidence intervals for any DWI test in Minnesota (blood, breath or urine).

Despite the 2009 National Academy of Science Report to Congress which requires that each lab analyze and establish confidence intervals for each type of test, Minnesota has failed. Nationally renowned attorney, Ted Vosk, has advocated with some success that no test result should be admitted as evidence unless it meets these criteria. (Note to practicing attorneys, judges and scientists: I highly recommend you read Mr. Vosk's work along with the NAS report to get a better understanding. See others including Justin McShane.)

How can I use measurement uncertainty in my case?

One well respected scientist has analyzed the data from Minnesota's Intoxilyzer 5000 to establish a confident interval. Analyzing test data provided by the Minnesota BCA of more than 44,000 breath tests, Rod Gullberg determined that the uncertainty of measurement of a .08 breath test in Minnesota's Intoxilyzer 5000 EN is approximately nine percent. Rod G. Gullberg, Breath Alcohol Measurement Variability Assessment with Different Instruments & Protocols, 131 Forensic Science International 30 (2003).

This percentage increases with the level of alcohol. At an alcohol concentration of 0.20, the measurement uncertainty is even greater. One must have a thorough understanding of complicated mathematical and scientific concepts to use his formula. For more information, see his work at the Borkenstein Course.

So what does this mean for drivers with alcohol concentrations close to critical levels? The Minnesota legislature has mad .04, .08, .16, .20 critical levels of alcohol concentration. In addition to other defenses, measurement uncertainty should be employed to defeat blood, breath and urine testing.