Critical Examination of the “Margin of Error” and “Uncertainty” in Breath Alcohol Testing

In every breath alcohol concentration (BAC) test, there is a great “margin of error” or “uncertainty”. This margin of error, or uncertainty, is a documented scientific fact.[1]

Every client I have represented who had a breath test, has had a margin of error in that breath testing that potentially worked in his/her favor.

In every DUI trial, there arises that point when the prosecution’s expert tries to relate the breath alcohol concentration (BAC) test result to the time of driving. It is key to the prosecution in any DUI case, to establish what the individual’s BAC was at the time of driving. It is essential to establish this in order to establish the crime. The BAC test is taken anywhere from 30 minutes to several hours after the time of driving. Therefore, to determine what the BAC was at the time of driving, the prosecution’s expert must engage in a process known as Retrograde Extrapolation, or as I like to refer to it, “voodoo science”.

It is imperative to the prosecution’s case that the prosecution’s expert establish what the BAC was at the time of driving. The prosecution’s expert will attempt to do this through testimony by estimating the number of drinks consumed by the defendant and testifying as to the BAC of the defendant at the time of driving.

It has long been known since the early 1930's, when Mr. Eric M. P. Widmark developed the first equations to calculate a blood alcohol concentration. Mr. Widmark determined that a degree of “uncertainty” exists in the establishing of a breath alcohol concentration. Mr. Widmark found two areas of “uncertainty”. He referred to one as the “rho factor” (food content in the stomach), and the other as the “B slope” (the burnoff or alcohol elimination rate).

Just recently, a new peer reviewed article has appeared expanding the number of areas of “uncertainty”. In his latest, and what I believe to be his seminal work, Rod G. Gullberg has raised the uncertainty issue to new level.[2]

There are really eight factors of uncertainty. They are:

1. Gender (male or female)
2. Weight
3. Rho factor
4. BAC at time of test
5. B slope
6. Time
7. Amount of alcohol in each drink
8. Amount of drinks consumed

Under the latest peer review article by Rod G. Gullberg, it has been determined that the estimation of the number of drinks, could be off as much as ± 25%. The estimation of the BAC at the time of driving could be off ± 42%.

These numbers show that breath alcohol concentration testing is at best a guess and at it’s very worst, voodoo science.

In his article, Mr. Gullberg points out that juries are often asked to consider and weigh a blood alcohol concentration. However, they are lacking appreciation of the uncertainty of these estimates. Juries tend to assign an unmerited amount of weight to the prosecution’s expert. In his conclusion, Mr. Gullberg stated that all forensic scientists should present when testifying, a reliable estimate of uncertainty along with their opinions as to any estimates of a BAC or an estimate of the number of drinks. For example, when a state expert renders an opinion that a person had five to six drinks, he should also admit to the jury that he could be off as much as 25%, which would mean the person had as little as three drinks or possibly as many as eight drinks. THIS IS JUST A GUESS.

When a state forensic expert is testifying as to the BAC at the time of driving, and assuming he renders an opinion that the individual had a .11 BAC at the time of driving, he should also admit to the jury that taking in the uncertainty into consideration, using an error factor of .42%, that would mean that the individual had as low a BAC as a .06 at the time of driving or a BAC as high as a .16 at the time of driving.

To put the significance of this uncertainty factor into laymen’s terms, imagine you are flying an airplane from Los Angeles to New York City. Assume at the half way point in your trip, the pilot comes on the speaker and says, “We’ll be on time to land today, but we might miss New York City by a few thousand miles ending up in possibly Maine or Virginia.” Would you want to fly on that plane or on that airline ever again? The answer is no. Then why should a jury believe a breath alcohol concentration test result that can be off by ±42%?

If the state expert does not in his direct testimony, give this uncertainty, then it certainly incumbent on defense counsel to draw these numbers out of him and at the conclusion of the case, I’m sure that the defense counsel would say to the jury as I would say, “Ladies and gentle men, you’re not dealing with true science as being presented here today, you’re dealing with voodoo. When you have a scientific estimate, it’s expected to be scientific-- not a guess!"

[1]Uncertainty Analysis for Forensic Science: Raymond M. Brach and Patrick F. Dunn; Lawyers and and Judges Publishing Company, Inc. 2004

[2]“Estimating the Uncertainty Associated with Widmark’s Equations as Commonly Applied in Forensic Science”; Rod G. Gullberg, Journal of Forensic Science International, 172 (2007) pages 33-39.

Copyright © 2007 Daryl B. Thompson, A Professional Law Corporation - All Rights Reserved