How DWI lawyers can expose the charade of the prosecution’s expert.
By Donald J. Bartell
Excerpted from Attacking and Defending Drunk Driving Tests
 Widmark Calculations and Presumption That Defendant Is Guilty
 Attacking Widmark Calculations
 Widmark Calculations in Breath Cases
The secret to winning any case is to first identify and then solve the predominant problem in the case. Only then do jurors begin to become receptive to what the lawyer considers to be the real issues in the case.
Lawyers think in terms of proving their case. For example, a personal injury lawyer wants to prove the plaintiff’s cause of action, a prosecutor wants to establish the corpus delecti, and a defense attorney wants to raise reasonable doubt. These are important, but what the lawyer really needs to be concerned with is how is he or she going to solve the predominant problem in the case. Until the predominant problem in the case is identified and solved (or at least lessened in impact), jurors will remain resistant to the evidence presented by the lawyer.
The Predominant Problem in Drunk Driving Cases
§10:30 Presumption That the Defendant Is Guilty
You can identify the predominant problem for the defense in a drunk driving case simply by talking to someone who is not a lawyer. Try this experiment: talk to five laypersons and tell them you are about to start a trial defending someone accused of drunk driving. See if there is a common concern.
Now reflect back on the last five pretrial conferences you had with the judges in drunk driving cases. Note the concerns the jurists had in those conferences. They parallel the concerns expressed by the laypersons.
Simply stated, most prospective jurors and judges are going to be of the opinion that the result in a drunk driving case is a forgone conclusion. And that conclusion is that your client is guilty. This is the predominant problem in defending drunk driving cases.
Jurors may even enunciate in voir dire that the case must be “cut and dried.” These jurors are certain that there will be a blood or breath test confirming their belief that the outcome of the trial is a foregone conclusion.
Similarly, the judge anticipates an early guilty plea by the defendant, and this stalls effective plea negotiations with the prosecutor. Judges will often ask the defense attorney what the defense is going to be. Note that the judge does not ask the prosecutor how the prosecutor is possibly going to prove the state’s case. Judges ask defense counsel this question because many of them cannot believe that you are serious about spending the next three days in trial on a .15% drunk driving case.
Practice Tip:
When you are invited into chambers before a drunk driving trial be prepared to respond to the judge’s inquiry about what your defense is going to be. On the one hand, you do not want to reveal the theory of your defense to the prosecutor. On the other hand, failing to respond to the court’s question is not an easy option. Be prepared with a truthful, but general answer. One way to lessen the risk of the court trying to ferret out of you the defense is by pretrial motions. Intelligent pretrial motion work will impress the court that you are not wasting the court’s time. Thus, your actual conduct in the case will have preceded you before the pretrial conference. If the court presses you for the specifics of your defense, you must ultimately inform the court that you cannot answer the question with the prosecutor present. This is difficult because essentially what you do is tell the court that the court’s inquiry was inappropriate. Nonetheless, it must be done. The goal is to have the prosecutor piece together the defense as close as to the time of the defense’s closing argument as possible.
§10:31 Widmark Calculations and the Predominant Problem
The more the defense lawyer can shake jurors loose from the predisposition of guilt (the predominant problem in a drunk driving trial) the more favorable the trial prospects of the defense.
Attacking Widmark calculations is an effective way to accomplish this mission because the Widmark formula uses averages to arrive at conclusions. Most jurors understand the problems inherent with using averages, and averages are the antithesis of proving something beyond a reasonable doubt.
When jurors hear that there is going to be a calculation using a long established formula, most jurors presume that the calculation is going to have some type of mathematical certainty.
However, when the jurors hear that in reality the Widmark calculations are actually nothing more than an educated guess, hopefully they will begin to question some of their initial assumptions. After the crossexamination on Widmark calculations many jurors may not be as predisposed to convict as when they were first seated. The defense campaign from the start is to get jurors to think “there is more to this than I thought—the case is not cut and dried.”
One way to remind jurors about the problems with averages that are used in Widmark calculations is to give them analogies that emphasize the problem with using averages. There are an endless amount of examples. For instance, you might tell jurors if one leg on a man’s pants is too short, and one leg is too long, the prosecution’s tailor would have you believe that on average the pants fit just fine. Or, if when you first turn the shower on in the morning, the water is biting cold, and then it turns to scalding hot, don’t worry, on average according to the prosecution’s plumber you are having a pleasant shower.
It is difficult to describe, but you will sense a perceptible change in the courtroom when the presumption of guilt starts to fade. Jurors seem more attuned to your questioning; the judge seems to give you more leeway in your crossexamination; the prosecutor seems a little more on edge. The difference is that the predominant problem in the case is beginning to become solved. This is the secret to winning drunk driving trials.
Attacking the Calculations
§10:40 Typical Presentation of Widmark Calculations
In a typical drunk driving prosecution, at some point the state’s expert will give an opinion as to how much the defendant had to drink. The prosecutor will ask the expert to assume a certain time of driving, a weight of the defendant, and give the expert the chemical test result.
The expert usually takes out a calculator and performs some intricate calculations. It looks serious, and with the expert’s deepstudied concentration while performing the calculations, it is an impressive show to the jury. There is some suspense while the jury waits for the result.
The result is frequently at odds with what the defendant told the police officer he or she had to drink on the night in question. The prosecutor then plays this point up to the jury.
Note, however, that usually nothing more is offered in the way of how this result was actually obtained.
§10:41 Expose the Charade
With the aid of a calculator, an expert has determined that your client had a high BAC. This conclusion further cements the jurors’ initial impressions that the defendant is in fact guilty. After all, an expert for the state has just as much as said so.
However, the way that these Widmark computations are presented provides a unique opportunity for the defense.
Ask the state’s expert if he or she discussed the case with the prosecutor before testifying. The answer is almost always yes. Then ask if the two of them discussed the calculations determining the number of drinks the defendant consumed. Again the answer is usually yes. Now draw on the fact that the expert knew the answer to the prosecutor’s drink calculation question before it was asked. Do this by asking (telling) the expert that since the expert already knew the answer to the question before it was asked, the drama staged by taking out the calculator was nothing more than an unnecessary charade for the jury. The following is a sample crossexamination:
Q: Did you discuss this case with the prosecutor before coming to court?
A: Yes I did.
Q: And as part of that preparation did the two of you discuss how much you believed the defendant had to drink?
A: Yes.
Q: You determined that by performing the calculations you just did for us in court today?
A: I used the same method.
Q: Then I take it prior to testifying today you knew the answer to the question as to how much you believed the defendant had to drink?
A: Yes.
Q: So taking your calculator out and spending all that time performing computations was not necessary?
A: Well, I like to be accurate.
Q: You knew the answer already—your computation was nothing more than a charade you performed for the jury?
At this point you really do not care what the answer is. The witness and the prosecutor have been exposed. If you are lucky, someone seated nearby in the jury box might begin to start thinking, if the case is so “cut and dried,” why is the prosecution going to such lengths to deceive us. The predominant problem for the defense is beginning to become solved.
There is not much downside risk in this line of questioning. If the witness denies going over the calculations with the prosecutor before testifying, you can simply move on to other questions. No one will know the significance of what you were about to ask, but did not. Also, the jury may think it is a little sloppy for the prosecution to put a witness on the stand without discussing the case with the witness first.
If you are still concerned about venturing into this area without more assurance of a positive reply, then watch the prosecutor and the expert during any recess. It is a common practice to have a recess before the prosecutor puts on the state’s expert. Generally, the police officer is the first witness to testify, followed by the expert. After the officer’s testimony it is a natural time for the court to take a recess. Some prosecutors even ask for the recess so they can talk with their expert. You can sometimes see the expert and the prosecutor doing the computations with the calculator in plain view in the hallway or at the counsel table during the recess.
§10:42 Ask What the Formula Is
Many experts do not really have an expert understanding of Widmark’s work. What they count on is that you do not have much of an understanding of the work either. They presume that they can get away with making Widmark calculations without discussing the limitations in the calculations, and usually they do.
You can expose imposters posing as experts on Widmark. Ask the expert if there was some formula the expert used to calculate the amount of drinks the expert claims the defendant drank. The expert will reply that there is such a formula. Ask if the formula used was the Widmark formula. The expert will say yes, or will say a modified version of the formula based on Widmark’s work was used. Either way the inquiry proceeds in the same manner.
Next ask the expert what the Widmark formula is. Many experts will reply with a formula that is not the actual formula. These experts, instead use a shorthand version of the formula in court. This shorthand version of the formula does not accommodate any of the variables that are in the actual formula. [For the actual formula, see Chapter 9, Blood Alcohol Concentration and Widmark Calculations.]
When the expert gives a reply with a formula that is not Widmark’s formula, ask the expert “this is not the actual formula is it?” The expert will then realize that you have more information on this topic than he or she expected. Only now will this expert own up to the fact that the version of the formula they gave is not actually Widmark’s formula. Finally, ask the expert what the actual Widmark equation is. Stunningly, the answer is often that the expert does not know.
You can emphasize in closing argument that the expert made calculations without knowing the actual formula.
§10:43 Ask Whether the Expert Has Read Widmark
Most experts have not read Widmark’s work.
If you ask an expert if he or she has read Widmark’s work, they may reply smugly that they have not read it because the work is in German. However, do not let the expert escape so easily. There is an excellent English translation of the work by Randy C. Baselt, Ph.D.
Prior to actually asking the expert if he or she has even read Widmark’s actual work, ask a series of questions from the work. Let the expert answer “I do not know” a few times before asking the ultimate question: “Have you even read the book?” This will help maximize the impact of their failure to read the definitive work.
If the expert does not know the actual formula, and has not even read the book, the strength of the expert’s opinions and calculations will be diminished. The following is a sample crossexamination of a criminalist’s knowledge of Widmark’s work:
Q: These calculations you did to determine the amount of drinks the defendant consumed, did you use some type of formula?
Comment: Always use your client’s name; the term defendant is used here for ease of reference only.
A: Yes.
Q: Is that the Widmark formula?
A: Yes.
Q: Well, could you please tell us the Widmark formula?
A: I would be happy to. The formula is the number of drinks, times 3.75, divided by your client’s weight, minus .02% an hour.
Comment:Expert’s frequently use a burnoff rate of .02% an hour for no other reason than it makes calculations in court easier. The actual average rate is lower, although, the normal range extends beyond this point as well.
Q: Is that the actual formula?
A: No, it’s a version I use for court. It is quicker so we don’t waste the jury’s time.
Comment:The witness will sometimes offer up some diversion hoping to avoid the next question. The examiner should simply make a mental note that the crossexamination appears to be working. Then plow straight ahead with your questioning.
Q: Well what is the actual formula?
A: I do not know. This is the one we use.
Comment:The witness realizes he or she has been caught and tries to drag others down with the witness by using the word “we.”
Q: Well, let me ask you this. You told the prosecutor on direct examination that the odor of an alcoholic beverage was one of the factors you used to base your belief that the defendant was impaired.
A: It was one factor.
Q: Didn’t Doctor Widmark indicate in his book that the odor of an alcoholic beverage is of a little assistance in determining if a person is under the influence?
A: I would agree with that.
Comment: This is a dodge to avoid answering the question of whether or not the expert knows what is in Doctor Widmark’s book. The witness would be better off acknowledging right away that he has not read the book. But the witness is still hopeful the question will never be presented.
Q: Do you know what Widmark found the coefficient of variation to be for men concerning the constant r factor?
A: No.
Comment: The answer is 13. Again, if the witness had answered on his or her own that the witness had not read the book, the questions would have already ceased. Failure to do so prolongs the agony.
Q: Well how about for women, what was the coefficient of variation for the r factor for them?
A: I do not know.
Comment: The answer is 10.
Q: Well, let me ask you this, how many men and women did Widmark use in his chapter on study of the kinetics of the conversion of alcohol, to determine the r factor?
A: I do not know.
Comment: The answer is 20 men and 10 women.
Q: Did Doctor Widmark find any correlation in his work between the constant ß and the constant r?
A: I do not know.
Comment: The answer is he did not find any correlation.
Q: You will forgive me, but have you read Widmark’s work?
A: No, it is written in German.
Q: Well have you read the English translation by Randy Baselt, Ph.D.?
A: No.
Q: You have not read any version of Widmark’s original work?
A: No.
§10:44 The Widmark Variables
There are two variables in the Widmark formulA: the ß factor and the r factor. [For an explanation of these variables, see Chapter 9, Blood Alcohol Concentration and Widmark Calculations.]
The important thing for lawyers to know about variables is that the numbers that can be placed in them are, by definition, variable. That is, they are not constant for all people. However, the shorthand formulas that the prosecution’s experts typically use assume constant variables for all people. These formulas assume that the defendant has an average r factor, and an average ß factor. In essence, the prosecution criminalist makes an educated guess about the defendant’s drinking.
This would not be so bad if the prosecution’s expert acknowledged that his or her calculations were based on assumptions. However, this is rarely done, and when it is done it is given only a passing reference. It is thus up to the defense attorney to explain to the jury the whole truth behind these calculations.
Practice Tip:
If the prosecution fails to explain the assumptions in their Widmark calculations, be sure to remind the jury about this in closing argument. Why did they not tell you they were making certain assumptions based on averages? Why did the prosecution not tell you that their use of a calculator was nothing more than a charade for you? Their expert knew the answer all along. And so do we. The answer is that the defendant is not guilty.
§10:45 CrossExamining on the Widmark Variables
The following is an illustration of how to exploit the Widmark variables in the crossexamination of the state’s expert. Note how the crossexamination ends. If you feel that you can get to this point with the expert in your trial, it is a perfect way to end the entire crossexamination of the expert, regardless of the type of chemical test that was taken.
Q: You previously explained to us that there is a formula that you used to calculate the amount the defendant had to drink?
A: Yes.
Q: That formula, does it have some variables in it?
A: Well, the formula I used contains constants in the formula in place of the variables. It has variables, but we use constants.
Q: I see, and by constants you mean averages?
A: Yes.
Q: Now one of these averages you used was the ß factor in the formula, something we sometimes call the burnoff factor?
A: Yes.
Q: You used a .02% burnoff factor for the defendant?
A: Yes.
Q: Well, that is not even the average is it?
A: No, but it is quite close.
Q: If the defendant had a burnoff factor that was higher or lower than average would that affect your ultimate calculation?
A: Somewhat.
Q: Well, what was the defendant’s ß factor on the night in question?
Comment: This is an opportune time to walk over to your client and lightly put a hand on the client’s shoulder while asking the question. It shows the jury in a subtle way that you are comfortable with your client, and conveys the point that your client is not some criminal. It is also good staging. The visual picture of walking over to your client makes the point of the question more memorable. Keep in mind that there will be a lot of facts introduced in the trial. No juror is going to remember all of them. Help the jurors remember the points you want to make with various types of emphasis. Voice inflection is one technique. Pausing after a key answer is another. Visual pictures, like walking over to your client while asking a question is a third method. Use them all.
A: I don’t know.
Q: There is another variable in the equation as well isn’t there?
A: I think you are referring to the r factor.
Q: And this r factor it can range from .5 to 1.0?
A: I don’t think it’s that big of a range. Widmark found the average to be .68.
Q: You did not actually read that in his work though did you? [Remind the jurors of the witness’ earlier deception.]
A: Like I said I have not read the work, but I have seen papers on the subject.
Q: Well, did not Lewis publish a study in one of these papers, “The Individual and the Estimation of his Blood Alcohol Concentration from Intake with Particular Reference to the Hipflask Drink” in the Forensic Science Society that the range was from .5 to 1.0?
A: I am not familiar with the article.
Q: You do agree that there is a wide range for men and women?
A: There is a range. I used the average.
Q: If the defendant’s r factor is different from the average would that also affect your calculations?
A: Yes.
Q: Well what is the defendant’s r factor?
Comment: Again, walk over to your client.
A: I don’t know.
Q: Well, if you do not know what the defendant’s r factor is, and if you do not know what the defendant’s ß factor is, you cannot tell us what the defendant had to drink on the night he was arrested, can you?
A: I can if I make certain assumptions.
Q: And those assumptions are that he is the average man?
A: Yes.
Q: And you don’t know if he is the average man, do you?
A: No.
Q: Without that assumption, you cannot tell this jury what he had to drink?
A: True.
Comment: Achieving this final concession does occur in actual practice. The chances of obtaining the complete concession depend on how much the witness has been battered in your crossexamination. If the witness has held up well, the witness is more likely to resist this concluding question by adding some qualifying proviso such as, “I cannot say for certain, but I can give you a good approximation.” On the other hand, if you sense a certain resignation of the expert, because the witness has been trumped one too many times, it is a good idea to shoot for this conclusion as a final point in your crossexamination.
§10:46 Checklist for Attacking Widmark Calculations

Hire an expert. If you do not think you need to hire an expert reread Chapter 9, Blood Alcohol Concentration and Widmark Calculations.

Focus on solving the predominant problem in the case. Everything should be coordinated toward reaching this goal.

Know Widmark’s formula. Have it in front of you when crossexamining the expert.

Know the averages used for the variables in Widmark’s equation. The average r factor is .68 for males, and .55 for females. The average ß factor is 0.15% per hour for men and women. Some more recent studies have found the ß factor to be .17% to .18% per hour.

Know the possible ranges for the variables in the Widmark formula. The ß factor ranges from .10% per hour to .25% pre hour. The r factor ranges from .5 to 1.0 for men (.60 to .87 is more commonly reported) to .44 to .80 for women.

Be prepared to expose the charade. Watch for the set up of the charade during court recesses.

In breath cases, look to see if you can get into partition ratios because of the claims made by prosecution’s expert about the defendant’s credibility by using Widmark calculations. [For more on this, see §10:50.]

Have some examples to give to the jury on why using averages to convict someone is inherently suspect. Make up your own. They are fun and easy to create.
Widmark Calculations in Breath Cases
§10:50 Calculations Attack Defendant’s Veracity
Widmark calculations take on special concerns when they are based on a breath machine rather than a blood test.
Typically, the prosecution’s expert opines that the defendant had more alcohol than the defendant told the police officer because the chemical test showed a higher alcohol concentration than the defendant reported.
The alcohol concentration measurement is based on what the defendant reported is obtained by using Widmark calculations. For example, the Widmark calculations may show that based upon what the defendant said he or she had to drink, the defendant’s blood alcohol level was .07 percent. However, a blood test may reveal a .12 percent blood alcohol concentration. The prosecutor’s expert then indicates that what the defendant told the police officer he or she had to drink could not have been true.
§10:51 Breath Machine Assumes a Partition Ratio
Breath machines assume that the concentration of alcohol in a person’s blood is 2100 times greater than it is in a person’s breath. It uses this relationship (known as a partition ratio) in calculating the blood alcohol concentration from a breath test.
Like any average, using one value can overstate or understate the true result. To take an extreme example, if a person’s partition ratio is 1050 to 1, instead of 2100 to 1, the breath test result will overstate the actual blood alcohol level by 50 percent. In this circumstance, the disparity in the results is not due to a misstatement by the defendant, but rather by the builtin error in the breath machine.
[For more on the partition ratio, see Chapter 11, Breath Testing.]
§10:52 Right to Attack Breath Machine’s Accuracy
In a breath case, if the state’s expert states that the defendant is not telling the truth based on Widmark calculations, the defense should consider arguing that the prosecution has opened the proverbial door to tell the jury the truth about breath machines and partition ratios. The defendant has a constitutional right to crossexamine witnesses. The witness is calling the defendant a liar, but it may be that it is really the machine that is lying. [For more on the right to crossexamine, see Pointer v. Texas, 380 U.S.400, 404 (1965) (defendant charged with robbery; court extended Sixth Amendment to the states). For how to introduce the variability of partition ratios into trial, see Chapter 12, Attacking Breath Tests.]
Many states preclude the introduction of evidence of the variance in individual partition ratios in breath cases notwithstanding the fact that such variances are universally accepted as true. The reason for this is that the legislature has defined the partition ratio as 2100 to 1. In effect, the legislature has ruled science out of the equation, and courts have upheld their right to do this. [See, e.g., People v. Capporelli, 502 N.E.2d 11, 15 (Ill. App. 1986) (constitutional challenge of statute using fixed ratio of 2100 to 1 denied; court stated that it cannot be said the legislature’s provisions are irrational); see also People v. Bransford, 8 Cal. 4th 885, 893 (California 1994) (consolidated drunk driving appeals denied defendants the right to introduce evidence that their individual partition ratios were different than 2100 to 1; the Supreme Court stated that the trial court correctly ruled such evidence is irrelevant and therefore inadmissible).]
Nevertheless, the right of crossexamination is still afforded an accused in a criminal trial. The value of crossexamination in exposing falsehood and bringing out the truth in trial has long been recognized. [See, e.g., 5 Wigmore, Evidence 1367 (3d ed. 1940).]
The Sixth Amendment right of confrontation is a fundamental right essential to a fair trial in a criminal prosecution. The United States Supreme Court has constantly emphasized the necessity for crossexamination as a protection for defendants in criminal cases. [See, e.g., Kirby v. United States, 174 U.S. 47, 55, 56, (referring to the right of confrontation as “[o]ne of the fundamental guarantees of life and liberty,” and “a right long deemed so essential for the due protection of life and liberty” that it is guarded against legislative and judicial action by provisions in the Constitution of the United States and in the constitutions of most if not of all the States composing the Union).]
What could afford more protection to the accused then the right to tell the jury the truth that the defendant was not lying when he told the officer what he or she had to drink on the night in question? The defense should argue that to protect the defendant from false claims made by the prosecution’s expert about the defendant’s veracity, the defendant has a federal constitutional right to tell the jury the truth about breath machines, partition ratios and Widmark calculations. If in a breath case the prosecution’s expert says that based upon the expert’s calculations the defendant must by lying, the door to partition ratios has been thrust open. [For more on attacking breath tests, and introducing partition ratios into trial, see Chapter 12, Attacking and Defending Breath Tests.]
Practice Tip:
Federalize the argument. You are asking the judge to tread into an area where the judge may feel a little uncomfortable. Provide the judge with some United States Constitutional backing.
Donald Bartell is on the Board of Directors of the California DUI Lawyers Association, and is a frequent lecturer around the state on DUI trial tactics. He has been asked to participate in the California DUI Lawyers Association and National Collegefor DUI Defense’s jury research project investigating what arguments resonate with jurors in drunk driving cases. He is the author of Attacking and Defending Drunk Driving Tests, from which this article is excerpted.