Avoiding Quicksand in the A swamp of Science

A back-to-basics approach to cross-examination of the state’s breath test expert.


We Have Met the Enemy… and He is Us 1

It’s hard to believer it’s been 15 years since the BAC DataMaster made its debut in Washington as the new state-of-art breath testing machine. In the first few years, challenges to the machine were so successful that few breath tests results ever reached the jury, leaving the Washington State Patrol Breath Test Section with a lot of spare time on its hands. Rather than playing Maytag repairmen, Director Rod Gullberg made good use of the time by preparing for the inevitable change of tide. He and Trooper Tiny McElroy made a religion of anticipating and trying to eliminate every defense challenger they’d ever faced-with admirable success. The defendant’s diabetic? No problem; we’ll do some “studies” proving it won’t affect the results. You don’t like one test? Fine, we’ll make the defendant blow twice. Paint fumes? Inhalers? Binaca? Their “experiments” on all that stuff and more proved just about nothing affects the accuracy of the Data Master test results.

In addition to the Breath Test Section’s efforts, the State Toxicologist (working hand-in-hand with prosecutors) systematically eliminated many records traditionally relied on to impeach the results and amended the Washington Administrative Code in order to solve whatever new problems the defense dreamed up. To compound the defense lawyer’s task, the current crop breath test experts are truly professional witness: experienced troopers, well-trained and tutored to anticipate just about every question defense lawyers have asked in the past. The result is a witness who must be approach with the same caution and respect one would show a coiled pit viper. Experienced lawyers seek only to make a simple point or two, and to then move on without doing any damage to the defense case. My own experience in taking on a state’s expert-one much less dangerous that today’s techs-without the requisite caution and proper preparation is instructive.

Many years ago, the state’s breath test program was run by Dr.Vidmatas Raisys, a very likeable man, but ill suited to the job of state toxicologist. Dr. Raisys hated lawyers and courtrooms and, as witness, possessed the guile and cunning of, say Winnie the Pooh, He was completely genuine, but often clueless when it came to forensic breath testing.2 Despite lacking the sophistication pf [resent day breath techs, Dr. Raisys proved how easy it is for an unwary defense lawyer (unfortunately, me) to climb out on a limb when cross-examining an expert on a subject the lawyer knows just enough about to get in trouble.

The occasion was a trial in front of several lawyers who’d come to watch me tear the DataMaster to shreds. I had a client who suffered “borderline diabetes” and I knew this means his breath contained an increased level of acetone whenever his blood sugar was out of balance. I also knew that acetone and alcohol looked very similar chemically, and that Dr. Raisys knew this as well, since we’d talked about it on a prior occasion. Thus, I made the mistake of assuming we’d agree that the acetone would affect my client’s breath test results. The pertinent exchange (Edited in an effort to make me look good) went something like this:

Me: Now, Dr. Raisys, you testified on direct that this so-called “instrument” measures the alcohol in a person’s breath by use of infrared spectroscopy. That’s just a fancy name for a light bulb, isn’t it?

Raisys: Well, actually, it’s a very sophisticated, state-of-art analytical method relied on for many years by the scientific community to detect and measure the present of various chemical compounds in the medium of light.

Me: Yeah, okay. But he human breath has all sorts of chemical compounds on it, even if the person hasn’t been drinking any alcohol at all, doesn’t it?

Raisys: Yes, that’s true.

Me: Well, Dr. Raisys this machine can’t differentiate between many things of these things and alcohol, can it?

Raisys: Things? What do you mean by things?

Me: You know, chemical things, stuff that is chemical.

Raisys: You mean chemical compounds?

Me: Yeah that’s it.

Raisys: [Enthusiastically] Oh yes, it most certainly can! You see, the detector board is programmed to isolate the oscillations of light waves caused by the vibrations of the hydrogen-carbon bond on the alcohol mole at three-point-one-six-nine-six-three-two microns, which is the chemical signature of ethanol molecules as they absorb energy from the light source. In fact the instrument is designed to be very specific for ethanol in its measurement.

Me: Well what about acetone? I mean, acetone looks just like alcohol, er, ethanol, to the machine, doesn’t it?

Raisys: No, not really. Acetone is similar in chemical composition, but that was accounted for in the design of the machine. We incorporated an interferant detector to guard against falsely high readings due to the presence of acetone. If acetone is present in any significant amount, the instrument will detect it, subtract it from the result, and record it on the evidence ticket.

Me: Well, in fact, a lot of people have significant amounts of acetone on their breath, don’t they?

Raisys: No, not really.

Me: What about diabetics? They have a whole bunch of acetone on their breath, don’t they?

Raisys: Not necessarily.

Me: Well, what id a diabetic did have a bunch of acetone on his breath and he blew into your precious instrument and the detector didn’t work right, then the test would show he was drunk on alcohol when he actually wasn’t, wouldn’t it, Doctor Raisys?!

Raisys: Which detector?

Me: The other detector, you know, the other one. The one that does the acetone thing!

Raisys: You mean the interferant detector?

Me: Of course.

Raisys: No, I don’t think so. I don’t think it would have a significant effect on the result.

Me: Aha, you don’t THINK so! Well, why not, Doctor?

So, keep your cross as simple and technical as possible, make a point or two for summation, then stop.

Raisys: Because if a diabetic had enough acetone on his breath to significantly effect the test result, he had be uh, dead.

Me: Oh…hmmm. No further questions

I know what you’re thinking: “I’m not as big a moron as Hayne. That would never happen to me!” Well, I didn’t think I was that big moron, either. It happened because I didn’t follow the simple rules I’ve set out below for cross-examining an expert witness-breath techs included. I don’t want the same thing happen to you, so I’m going to suggest a technique for cross-examining the breath tech designed to avoid such a disaster.

Avoiding Quicksand in the Swamp of Science

It’s true, DUI cases are very difficult to try, and jurors do love the numbers. But all is not lost for the wise defense lawyer. Good cases are still good and bad cases are still bad. When the case has bad driving, bad physicians, bad attitude, and bad breath test-yes, you have a bad case. But in those more typical cases where the prosecution’s evidence is not a wall-to-wall defense disaster, reasons for doubting the breath test can still be found, but those doubts are in the rest of the case, not in attacking the breath tech.

The next obvious question is “Okay, the what do I ask the breath tech?” That’s an excellent question. Unfortunately, I can’t answer it. Any list of questions is good for about 24 hours, after which time the Breath Test Section will have developed answers that will devastate the questioner. My technique for cross minimizes the risk while maximizing the return. It comes down to a few simple rules based on many bad (and some good) experiences:

1. K.I.S.S (Keep it Simple, Stupid).

Anyone who has tried a DataMaster case will verify that the more technical the desired point is, the tougher it is to make-and due to the mind numbing convolution necessary to get to it, the less impact it will have on the jury. Remember (and keep reminding yourself) the average juror knows little or nothing about infrared spectroscopy or analytical chemistry and doesn’t much care. Jurors just want to know if there is some legitimate reason, any legitimate reason, to doubt the results in this particular case. So, keep your cross as simple and non-technical as possible, make a point or two for summation, then stop.

First, get the tech to agree to some simple ground rules:

Q: Trooper, I wasn’t very good in chemistry and don’t want to confuse the jury, so if I ask you a question you don’t completely understand, please ask me to try again, would you?

A: Alright.

Q: And I’ll try to keep my questions simple enough for a “yes” or “no” answer. But if I don’t, please ask me to rephrase it, so it can be answered “yes” or “no”. Okay?

A: Sure

Now, the trick is to make the tech stick to your agreement:

Q: Is the DataMaster machine designed to screen for any interferant besides acetone?

A: Well, actually, it doesn’t matter since…

Q: [Immediately] Excuse me Trooper. I screwed up again. I promise to ask only questions you could answer “yes” or “no.” Let me ask the question again. The DataMaster is not designed to screen any interferant besides acetone, is it? Do you understand the question?

A: Yes

Q: Are you able to answer it “yes” or “no”?

A: Yes

Q: And what is the answer?

A: No.

Next, try to avoid leading with your chin:

Wrong approach: Q: Now, the machine’s slope detector has a history of being unable to determine the presence of mouth alcohol, doesn’t it? A: A history? No. In fact, our studies have shown beyond a shadow of a doubt that the slope detector will abort the test if it detects any anomaly in the breath sample. While it’s true we’ve been able to trick it on a few occasions in the lab, our experience in the field in tens of thousands of tests has shown that the slope detector is extremely reliable as long the operator follows the testing protocol. In other words, as long as the officer administering the test kept your client under continuous observation for 15 minutes prior to the test, as I understand he did in this case, I’m confident the slope detector was operating properly and the test results are accurate and reliable.

Right approach:

Q: This machine isn’t perfect, is it?

A: No, no machine is.

Wrong approach:

Q: Trooper, you’ve conceded the breath tube must be hot enough to avoid condensation, but what if it wasn’t? What if alcohol was present in the tube from the drunk who blew into it a few minutes before my client?

A: Well, our studies have shown it wouldn’t have any effect, even if the tube were stone cold. Heating it really isn’t necessary, but we do all we can to benefit the defendant, to make sure that no test is completed if there is any question as to its accuracy. So, if there alcohol present in the tube, which is highly unlikely, the instrument would shut itself down since it would be unable to zero itself before the defendant provided a sample. We do all we can to give the defendant every benefit of the doubt.

Right approach:

Q: Trooper, you’ve conceded that the breath tube must be hot enough to avoid condensation, but isn’t true Officer Bully had no way to determine its actual temperature other than touching it?

A: That’s correct.

Q: And condensation can attract and trap alcohol molecules, can’t it?

A: Yes

2. Plan Ahead. Know where you’re going, why you are going there, how you are going to get there, and for heaven’s sake-when you’ve arrived. After deciding what concession you want tech to make, ask yourself “why?” Is it worth the effort? Once you’ve decided it is, you must spend a lot of time figuring out how to get there. You literally must ponder each question, asking how the witness could misinterpret or avoid it, backing up question by question until you’re confident there is only one possible answer for each one. In other words, ask yourself: If I ask the tech this question this way, is there any way he or she can avoid giving me the answer I want? If so, retreat and break the question down into smaller and smaller components until you’re satisfied there’s no way out. Then do the same thing for the next step, gradually building a fence from which there is no escape. Finally. Don’t get greedy. Making points with some of these witnesses is harder that pulling teeth! When you’ve made it, stop. Period. Otherwise, you may find your carefully crafted web unraveling before your eyes.

For example, let’s say you want to make the point that the machine’s software did not keep a record of “error codes” two and three (sample chamber temperature too low or high).3

Wrong approach:

Q: Trooper, does it make any difference that the machine doesn’t keep track of the sample chamber temperature error codes in the database?

This type of generalized question is very dangerous. It invites the state’s expert to make a speech that destroys the value of the point you’re trying to make (Absolutely not [Turning to jury with a friendly smile] Let me explain…”). The way to make the point is to start by answering the following questions:

  • Where do I want to go? I want the technician to admit the machine doesn’t keep a record in the database when the sample chamber temperature is either too low or too high (error codes two and three) because the software was not programmed to record them.
  • Why do I want to go there? To show the jury (1) the machine uses software that fails to keep track of vital information and (2) if the temperature in the sample chamber was in fact too high or too low near the time of my client’s test, it could effect the accuracy of the results and we wouldn’t even know it.
  • How am I going to get there? I must anticipate every way the witness can elude the point and must corral the technician with a series of simple, leading questions designed to make the desired end inevitable.
  • How will I know I’m there? When I have enough of a confession to argue the point in summation.

For example:

Q: Trooper Expert, isn’t it true this machine uses something called error codes to identify problems in its operation?

Q: And, in fact, those error codes are numbers which show up on something called the database?

Q: You recognize what I’m showing you, marked for identifications as defense exhibit one, as a typical database, don’t you?

Q: And it’s true, isn’t it, that a database such as this is permanent record of all the breath tests performed on a machine, right?

It can be long and difficult to force the tech to concede a point in your favor.

Q: And also contained in the database is a permanent record of the error codes we spoke of a moment ago, correct?

Q: For example, if one were to see an error code number six on the database, it would mean that a “fatal systems error” occurred in this machine, right?

Q: And a “fatal systems error” means there was critical failure in the computer software, doesn’t it?

Q: And isn’t it also true that the reason the machine uses this error code system is do that anyone interested -technician, prosecutors, defense counsel or even the jury, for instance -would know when there was a problem with the machine?

Q: And isn’t it also true that the various error codes are numbered one through twenty by the manufacturer?

Q: Again, for instance, a number six signifies the “fatal system error” we spoke of a few moments ago, right?

Q: Now, the error codes two and three have to do with the temperature in the sample chamber, don’t they?

Q: Number two means the temperature was too low and number three means it was too high, right?

Q: Now, the sample chamber is where my client’s breath was actually measured, isn’t it?

Q: And the temperature of the sample chamber is critical to the correct operation of the machine, isn’t it?

Q: And it’s also true, is it not that the machine used to test my client’s breath in this case was incapable of making any record in the temperature in the sample chamber was either too high or too low, isn’t it?

Q: So if, in fact, the sample chamber temperature was too high or too low near the time of my client’s breath test, it wouldn’t have been recorded anywhere, would it? 4

Now, are you there? Some would say yes, others no, since we haven’t established the effect of a high or low temperature on the machine’s accuracy. Personally, I think we’re close enough to argue the point in summation, which is my acid test. You can see from this relatively simple example how long and difficult it can be to force the tech to concede a point in your favor. That’s another reason to keep your goals modest; one or two points that are used in conjunction with the rest of the case to raise a reasonable doubt.

3. Lead, Lead, Lead. Hundreds of witnesses have taught me this lesson again and again. It is especially good advice when dealing with experts, since the way the question is asked can make a world of difference in the response. For example, let’s assume I want to show the jury that the “interferant detector” doesn’t always work properly. How am I going to get that concession?

Wrong approach:

Q: Trooper, does the interferant detector on this machine always work properly?

The result: even though it doesn’t, I’m sure to get a very qualified “no” along with a killer speech explaining why it doesn’t matter (‘the defendant would be dead’).

Right approach:

Q: Trooper, in the course of your training, you’ve certainly been made aware of the fact the there are many compounds on the human breath besides alcohol, haven’t you?

Q: And you were also made aware of the study that identified as many as 102 such compounds on healthy human subject, weren’t you?5

Q: Now, it’s true, is it not, that the manufacturer was concerned enough about this problem that it attempted to screen a subject’s breath to make sure the result five was actually a measurement of alcohol rather than something else?

Q: In doing so, the manufacturer built into the machine something known as the “interferant detector,” didn’t it?

Q: And it is also true that this components is designed to screen compounds on the breach that appear chemically similar to alcohol to the machine’s detector, such as acetone, isn’t it?

Q: Certainly you would agree with the manufacturer, that this component is critical to the correct operation of the machine?

Q: Now, it’s true, is it not, that you are aware of repeated instances in which the interferant detector has malfunctioned in DataMasters?

Q: Well, in fact you gave either personally observed or have been apprised of instances in which the interferant detector registered an interferant on one of the two breath samples, but not the other, haven’t you?

Q: And, in fact, you would certainly agree that if an interferant, of say 0016, appears on one of the person’s samples, you would expect it to appear on the other wouldn’t you?6

Q: So, on those occasions when that happened where the interferant level was well over the threshold, the interferant detector may not have been working properly, correct?7

Now, are you there? I think so, but you could ask the next question:

Q: And, in fact, if the interferant detector wasn’t working properly something other than alcohol could have been added to my client’s alcohol reading on the evidence ticket, couldn’t it?

These examples illustrate that getting to your point can be a laborious process, made even more difficult by an uncooperative technician. That’s why the rules of keeping it simple, knowing where you are going, why, how and when to stop, and the use of leading questions are so important. In the final analysis, the tech is going to win any challenge to the DataMaster test result that isn’t carefully thought out and planned with an almost paranoid precision. The embarrassment of my ineptitude with Dr. Raisys still burns after all these years. By diligently following these few simple rules, you can hopefully avoid following in my footsteps.

Steve Hayne is a partner in the Bellevue firm of Hayne, Fox & Bowman. He is a past president of WACDL.


1. Walt Kelly, Pogo, circa 1959

2. In fairness to Dr. Raisys, in those days the position of the state toxicologist was sort of an “honorary” appointment, part-time and under-funded. His real job was being a professor at the University of Washington

3. Please note: This example illustrative purposes only and does not apply to cur rent DataMaster Software.

4. Note the question is about the temperature “neat the time of the client’s test” rather than at the time. The state takes the position that the machine would shut down and abort the test if the temperature were too low or too high during a test sequence, since the “system” wouldn’t “zero,” i.e., verify that no alcohol was in the chamber. This is an example of how precise your questions must be to keep out of trouble.

5. Krotosyznski, et al in the Journal of Chromatographic Science, 15, 244 (1977), a well-known study with which the technician should have some familiarity.

6. If the tech wasn’t to play games with you, he or she will argue this point by claiming that the interferant may have been present “just enough” to trip the detector on one sample but just barely under the necessary level on the next. Thus, you should have examples of detections over that minimum level (0.1) which were detected on only one sample. Such example can be found by searching databases for various machines.

7. Notice the question is whether the interferant detector “may not have been” working properly rather that wasn’t. If you ask the later, the tech will doubtless disagree and go into a lengthy speech on the detector (0.1) and argue that on one sample the interferant level was slightly over that threshold and, on the other, slightly under it.