Our firm regularly teaches DUI defense law to other lawyers.  This page is an adapted and abridged version of the the draft for a chapter we wrote for a book titled "Effectively Defending DUI and Other Serious Traffic Offenses."  For a less technical, "layman" overview of DWI laws, click here.  For frequently asked questions, click here.

Albo & Oblon lawyer, David Oblon, pictured here, regularly lectures and writes on DWI defense law.  Here he talks about the science of DWI.

This page (and all the information on this web site) is copyrighted.  All rights reserved (2005).
           

                                                 Scientific Defenses

             A.            The Science of Driving Under the Influence of Alcohol.

             Unlike in most criminal cases, alcohol impairment cases often turn on scientific evidence.  The government supports arrests with field sobriety tests to determine probability of impairment.  These tests are distracted attention tests from which an officer extrapolates alcohol impairment.  It wins convictions based on complicated chemical breath or blood tests to determine the level of impairment.  Indeed, one of the elements of the offense – the blood alcohol level – is, itself, directly the result of scientific evidence.

             Practitioners must know the science and assumptions behind such tests in order to assure their veracity and identify flaws.  This chapter reviews the science of all these tests.

                         1.            Field Sobriety Testing.

             There are no mandatory standards governing field sobriety testing.  Officers receive training that includes field sobriety training, but it appears that such training is clumsy and haphazard at best.  Indeed, experienced practitioners will quickly learn that individual officers use a wide variety of tests different from one another – and they conduct each test rather differently.  This is true even among officers of the same department.  The most common tests administered are the “Walk and Turn” test, the “One Leg Stand” test, the “Nose Touch” test, “the Finger Touch” test, the “Alphabet” test, the “Numbers” test, the “Horizontal Gaze Nystagmus” (“HGN”) test, and the preliminary breath test.  Usually, an officer will offer three of these tests, plus the preliminary breath test.

            The National Highway Traffic Safety Administration (“NHTSA”) has studied and approved a battery of field sobriety tests.  Practitioners should use these tests as a baseline from which to critique different versions of the tests offered by an arresting officer.  The NHTSA battery consists of the “HGN” test, the “Walk and Turn” test, and the “One Leg Stand” test.  According to a study adopted by NTHSA in 1998, the combination of the three tests accurately predicts a blood alcohol content level of 0.08 or greater 91% of the time.  The study indicated that, used alone, the HGN test was accurate in such prediction 88% of the time, the “Walk and Turn” was accurate 79% of the time, and the “One Leg Stand” test was accurate 83% of the time.  See Development of a Standardized Field Sobriety Test, Appendix A, NHTSA (available online at http://www.nhtsa.dot.gov/people/injury/alcohol/SFST/appendix_a.htm) (hereinafter “NHTSA Study”).

            There do not appear to be any studies confirming the veracity of the myriad of versions of this battery used by police in D.C.  This results in numerous defenses to the tests.  One should think about flaws in a particular test and point these out to the court.  In every instance, the tests should be “expanded” so that the judge balances the good along with the bad.  For example, if the officer testifies that, on the “Stand on One Foot” test, that the accused put his or her foot down at count number 21, point out that he or she successfully kept his foot up for count numbers 1 to 21 and 21 to 30, did not sway, did not hop, did not use arms for balance, did not miscount, did not forget to count by thousands . . . Don’t stop until the judge tells you to shut up! 
 
                                    a.            Blanket evidentiary objection.

             No matter which test an officer uses, one should never permit the officer to testify as to the correlation between an accused’s performance on a test with a level of alcohol impairment.  While it is acceptable for an officer to testify that he instructed an accused to stand on one foot for 30 seconds and to further testify to what the officer saw, it is improper for an officer to summarily conclude that the accused performed the test “poorly,” or that he concluded as a result of the test that the accused was impaired by alcohol.

                                     b.            “Walk and Turn.”

             According to NHTSA, the accused is directed to take nine steps along a straight line, as if walking a tightrope.  After taking the steps, the accused must turn on one foot and return in the same manner in the opposite direction.  There are 8 “indicators” of impairment:  (1) the accused cannot keep balance while listening to the instructions, (2) the accused begins the test before the instructions are finished, (3) the accused stops while walking to regain balance, (4) the accused does not touch heel-to-toe, (5) the accused steps off the line, (6) the accused uses arms for balance, (7) the accused makes an improper turn, and (8) the accused takes an incorrect number of steps.  If the officer identifies two or more indicators, the studied reliability percentage applies.

            District of Columbia practitioners will commonly see this test offered with only an imaginary line.  Is this more difficult?  Without studies, it is difficult to tell.  Practitioners will also commonly see clients who were ordered to stand heel to toe while the instructions are delivered.  Standing in this unnatural pose for a minute while listening to important instructions from a police officer may be more challenging than intended.  It is very common for an officer to identify one (and only one) clue which, to the unsophisticated, would seem to indicate probable impairment.  For example, police commonly will say that an accused “stumbled on the turn.”  However, according to the NHTSA Study, this amounts to only one indicator of impairment and does not show that the accused is probably impaired.

                                     c.            “One Leg Stand”

             NHTSA prescribes that, for this test, an accused must stand on one foot with the other foot six inches off the ground.  He or she should also count by thousands to 30.  There are four “indicators” of impairment:  (1) the accused swayed for balance, (2) the accused used arms for balance, (3) the accused hopped to maintain balance, and (4) the accused put his or her foot down early.  The officer must identify two or more indicators to conclude that the accused is impaired.

            In the District, police will commonly conclude that one is impaired if he or she puts his or her foot down early.  However, this is but one indicator and should be insufficient for establishing probable cause.  Some D.C. police will not tell the accused that he or she should count to 30, rather the instruction is to count until the officer says to stop.  Does this uncertainty make the test more difficult?  

                                    d.            Horizontal Gaze Nystagmus.

            Horizontal Gaze Nystagmus (“HGN”) test measures the involuntary jerking of the eye when a person is impaired by alcohol.  According to NHTSA, normal, sober people have a nystagmus at high peripheral angles (i.e., when they are looking to the extreme left or right).  However, the studies seem to show that those who are impaired by alcohol have exaggerated jerkings that can be observed at less extreme angles, such as 45°.

             In conducting this test, the officer tells an accused to keep his or her head still and follow a pen that the officer waves horizontally across each eye.  The officer looks for three indicators of impairment:  (1) the eye cannot smoothly follow the pen; (2) there is distinct jerking at maximum deviation, and (3) the jerking’s angle of onset is within 45° of center.  If and officer can identify 4 or more clues, then the officer can conclude that the accused is probably impaired.

             This may be reliable in theory, but this test is facially absurd in practice.  Many judges will ignore HGN evidence at trial.  How does the officer really measure the angle?  What is the difference between a natural nystagmus and a “distinct jerking at maximum deviation?”  How can the officer eliminate the consumption of seizure medications, phencyclidine, a variety of inhalants, barbiturates, and other depressants that also cause a nystagmus?  How does the officer’s flashing overhead service lights affect this test?  Why does every officer who ever does this test always find all six indicators on every test?  If one appears before one of the few judges who take this test seriously, one should be prepared to cross-examine the officer on these items.  Argue to the judge that your client does not object to the officer testifying that he or she asked the client to follow the pen with his or her eyes; nor does your client object to the officer testifying that he or she saw a lack of smooth pursuit, a jerking at maximum deviation, or a jerking onset at 45°.  Rather, state that your client objects to the officer concluding that alcohol caused those reactions without expert witness testimony.  Of all the field sobriety tests, the HGN is the most rooted in scientific evidence. 

                                     e.            The preliminary breath test.

             The preliminary breath test (hereinafter, “PBT”) is somewhat of a quagmire.  If admitted and given weight, it alone can provide grounds for probable cause.  However, it is very difficult for the government to have the results admitted to evidence, so it is usually not a big issue.  

             The PBT is simply a hand-held breath test.  The D.C. Code bars the use of the PBT during a trial.  However, it is admissible in a pre-trial Motion to Suppress.  However, there are still a great number of defenses to the PBT.

             For example, the PBT does not have any evidentiary exceptions to basic rules of evidence as does the official test after arrest.   Therefore, a prosecutor should be unable to offer this scientific evidence without an expert witness.  Be aware, however, that many judges take the position that the rules of evidence are relaxed for Motions to Suppress and that PBT results can be offered in the context of such a motion.  It should be very difficult to lay an evidentiary foundation for the admission of a PBT.
  
                                    f.            The other tests.

             Since there is no identifiable standards or science to back up the many other tests administered by D.C. police, it is important to point out the flaws.  For example, on the Nose Touch” test, make sure it is clear that the officer told the accused to touch the “tip” of the nose as opposed to the “top.”  Ask the officer why some officers do this test with the accused’s eyes closed and others with the eyes open.  Ask why some ask for one touch each from the right and left arms, while others make a game of it asking for a right or left arm touch at the officer’s command, with the officer sometimes switching up the sequence.  On the “Alphabet” test, ask why the officer didn’t ask the accused to simply say the alphabet from A to Z, but insisted on an alphabet subset, such as from C to T?  Almost invariably, the officer will say that it was they way he was taught.  When pressed, the officer will concede that he did the subset to make the test more challenging. 

                         2.            The Intoxilyzer 5000.[4]

             The District of Columbia prefers breath tests -- although it can take blood or urine samples.  In practice blood is taken only in accident cases; urine is never taken.  

             Washington, D.C., has been using for some time the Intoxilyzer 5000 for breath testing.  Many other states use this device, too.  It is important to know both how the machine works and the proper operation procedure in order to spot testing irregularities, inaccuracies, and flaws.

                        a.            Scientific principles.

              The Intoxilyzer is an infrared device.  It works by passing an infrared light beam through a breath sample and measuring the amount of light absorbed by alcohol.  The more alcohol, the more light is absorbed, the higher the corresponding result.  

              The scientific theory behind this sort of test is based on the appropriately-named Beer-Lambert Law and Henry’s Law.  According to the Beer-Lambert law, if a light is directed through a container, the amount of light detected on the other side of the container is decreased by any substances in the container in proportion to their absorption coefficients and concentrations, and is also decreased in proportion to the distance across the container.  In English, this means that molecules absorb light.  (Anyone who stands outside on a hot, sunny day in a black shirt knows this).  If you shine a light through a box with “stuff” (molecules) in it, some of that “stuff” will absorb some of that light.  If you know how much light you directed at the “stuff” and can measure how much light comes through the other side, you can identify the amount of the “stuff.”

             All “stuff” (molecules) will absorb light at different wavelength patterns.  Thus, if you know the pattern for alcohol, and scientists claim that they do, and you see the light absorbed in the “alcohol” pattern, you can conclude that the “stuff” absorbing the light is alcohol.  The Intoxilyzer uses five filters to identify “ethyl” alcohol wave patterns, and is supposed to differentiate between ethyl alcohol and methyl, isopropyl, or butyl alcohols.

             According to Henry’s Law, the weight of any gas that dissolves in a definite volume of liquid is directly proportional to the vapor pressure that the gas exerts above the liquid.  In English, this means that a certain temperatures, there will be a definite ratio between the concentration of alcohol in lung air and in body fluids, such as blood.

             The Intoxilyzer combines these two scientific principles.  It takes a breath sample, which should contain a constant alcohol/liquid ratio with the subject’s alcohol/blood, shoots light through it, knowing how much light should be absorbed, looks for the alcohol wavelength “signature,” and translates the results via an algorithm to be micrograms per mililiter of breath.  This is commonly reported in "blood test" nomenclature of percent by weight by volume.

                         b.            Operating procedures.

             An operator begins a breath test by starting the machine and entering the correct PIN number.  The machine will confirm that the operator is licensed.  The operator will then some additional identification information.  

              The machine will then conduct a series of internal checks:  (1) a “Prom” check, which tests the validity of the computer program; (2) a “RAM” check, which tests for memory failures; (3) a “Temp” check, which tests the temperature of the sample chamber; (4) a “Processor” check to locate the computer processor; (5) a display of the software version and serial number; (6) a “Printer” check; a (7) a “RTC” check, to confirm the time and date of the computer; and (8) a standards test to test the internal standards.

              Assuming the checks are negative, the machine will then test the simulator temperature.  The result must be between 33.8° and 34.2° Celsius.  It will then conduct an “air blank,” which is essentially a breath test without anyone blowing into the machine.  If there is alcohol in the air, this air blank should detect it.

              The machine next runs, automatically, an artificial simulation to test itself.  Basically, it artificially simulates the breath sample of a person with an artificial breath sample that has a known alcohol concentration.  In the simulation, the machine pumps a pre-measured, heated solution of alcohol and water into the machine.  This solution is supposed to be between 0.090 and 0.105 grams per 210 liters of air.  The machine then conducts, essentially, a single-sample breath test on that artificial sample.  If the results are within the machine’s range of tolerance, then the machine assumes that it is working accurately.

              Next, the machine conducts a second air blank test and requests a breath sample from the accused.  The accused blows one time and the operator takes the breath tube off the device while the machine analyzes the sample.  The machine conducts a third air blank test, the operator puts the tube back on the machine and the accused blows a second time in to the machine.  Finally, the machine conducts a fourth air blank test and displays the result.  If the difference between the two samples exceeds 0.02, then the machine will request a third breath sample.  The reported result is the lower of the two (or three) breath samples with the third digit truncated (neither rounded up nor down).

              The Intoxilyzer is not a perfect device.  For example, it unfairly assumes that that human breath is 34° Celsius (93.2° Fahrenheit) when leaving the mouth.  At this temperature, it assumes that the ratio between the concentration of alcohol in the blood and in the lungs at equilibrium is 2100:1.  However, people are all different and if an individual’s temperature were higher than this assumption, then ratio should be different and the corresponding breath test result will show a higher alcohol concentration than the actual alcohol concentration.  If, for example, one has a fever, the Intoxilyzer result will be unnaturally high.  This problem could be abated if the police were to take the subject’s temperature prior to the test.  It also assumes that one does not have a measurable amount of endogenous alcohol.

             The device also has various error margins that could result in the reported test result being inaccurately high.  For example, the machine takes two samples for a completed test and has a tolerance of 0.02 between the samples.  This, then is the machine’s “official” margin of error.

             Some other sources of error could include the detection of substances other than alcohol.  Recall that the machine has five filter wheels for interfering substances, such as acetone and toluene, and other forms of alcohol.  These filter wheels are designed to detect interfering substances and, if the amounts exceed 0.01 grams per 210 liters of breath, it should cancel the test and report an “Interferant Detected.”  However, while the simulation test confirms that the machine detects alcohol, there is no test – ever – that confirms that the filter wheels are really working.  Also, if a substance for which there is no filter is detected, the machine will report the substance as “alcohol,” and incorrectly report a subject’s breath test result as higher than reality.

             The machine runs the risk of detecting alcohol that is concentrated in the mouth and not contained in deep lung air.  This is aptly called “mouth alcohol” and causes great concern.  Burps can affect the veracity of the test.   Recall that the Henry’s Law ratio of 2100:1, discussed above, assumes that the breath sample consists of deep lung air.  With a mouth air breath sample, this ratio would be different.

             All breath test operators should watch an accused for 20 minutes prior to the administration of the breath test.  This is to confirm that the accused has not ingested fluids, regurgitated, vomited, eaten, or smoked.  It is also to confirm that the subject has not burped or belched.  While the Intoxilyzer is supposed to recognize that mouth alcohol will cause a spike in breath reading that should trigger an automatic cancellation of the test, this is not fail-safe device.  Some expert witnesses (and some breath test operators) will tell you that if mouth alcohol is breathed at a steady rate into the machine, the machine will not notice a “spike” in the reading and may report an inaccurately high reading.  This is most problematic in cases where the accused has alcohol-impregnated food particles or other foreign objects in the mouth.

             There are potential problems with the simulation test.  Defense attorneys are never provided a copy of the certification the simulator solution.  This is very important because, if the solution is inaccurate and the test misidentifies the sample, then the machine is obviously broken.  By analogy, this is akin to a police speeding radar device that is tested by tuning forks.  Defense attorneys in speeding cases always check to see if the radar was properly calibrated and that the tuning forks used to test the device were also calibrated.  This important safeguard is absent in breath tests.

             It is always important to learn the results of both breath samples, the results of the simulator test, and the results of the test conducted back when the simulator solution was first used.  These are always different numbers.  If one adds up the discrepancy between the four numbers, one may be able to demonstrate that the reported breath test result is inaccurate. 

              Aside from error margins, there are operational defenses to the Intoxilyzer.  For example, if the accused does not blow enough air into the machine, the machine will cancel the test and report a “deficient sample.”  This is often confused with an “invalid sample,” which occurs when an accused has residual alcohol in his or her mouth.  Commonly, a breath test operator will see a reading of “invalid sample,” confuse it with “deficient sample,” and get mad at the accused for “not blowing hard enough.”  Many times, such an accused is charged with Refusal to Submit to a Chemical Test!  Certainly, these cases should be dismissed.

                                     c.            Statutory compliance issues.

             D.C. law sets forth a “fast track” evidentiary path for the admission of breath test results as an exception to the hearsay rule.  The United States Supreme Court has probably invalidated this scheme through its recent decision in Crawford v. Washington, 124 S. Ct. 1354, 2004 U.S. LEXIS 1838 (2004), which holds that the Sixth Amendment constitutional right to confront and cross examine witnesses trumps evidentiary rules.  The lower courts have not carved out a “DUI exception” to the Crawford rule as of this writing, but eventually either judicial or legislative action will “cure” this difficulty.  In this event, it remains necessary to know the D.C. statutory scheme for breath tests.

            Before a police officer even administers a breath test, the officer must advise an accused that he has already consented to the test under a theory of “implied consent.”  
   
              After the police inform an accused of the “implied consent law,” and the accused does not refuse, then the breath test begins.  

              The breath test operator must complete a checklist and fill in a log.  These items are available -- but not given in ordinary discovery.  

               A defense attorney, in order to challenge the breath test, must usually have the breath test operator present at trial.  In order to summon this officer, the lawyer must notify the government at least 10 days before the trial of this demand along with reasons why the officer is necessary.

               d.            “Relation back.”

            The breath test is rebuttable, even for the “per se” application to those accused with blood alcohol concentrations over 0.08.  However, the defendant needs to affirmatively supply the evidence for the rebuttal.  After all, the test results are presumed indicative of the blood alcohol level at the time of the alleged offense.  And, the argument that one’s “great” performance on field sobriety tests, by itself, rebuts the test is rarely successful.

            One of the best defenses is the argument that one’s blood alcohol concentration increased during the time it took for the officer to seize an accused and test him or her.  Due to the nature of how the body processes alcohol, it is possible to have alcohol in the body that has not yet been absorbed into the blood.  As a result, one with unabsorbed alcohol in the body will have his or her blood alcohol concentration increase over time even though he or she has stopped consuming more alcohol.  If the person’s blood alcohol concentration was below 0.08 at the time he or she was driving, but rose over this level by the time he or she was tested, there may not be a violation of law.  An expert witness can review an accused’s body size, the time and types of alcohol he or she consumed, and the time and types of food he or she consumed.  Using this information, the expert can determine the blood alcohol level of the accused at various points in time.  However, be aware that, most of the time, an accused’s blood alcohol concentration is declining not rising.

                        3.            Blood tests.

             Ever since the accused lost the right to choose between blood and breath tests, the number of blood test cases has declined.  Nonetheless, they are used in accident cases.  The procedural requirements are considerably more extensive than those for breath tests, which makes blood cases more of a challenge than breath cases.  For example, counsel must subpoena the person who did the blood analysis and, along with an expert witness, cross examine the lab technician to test the scientific reliability of the blood test and its methods.

                        4.            Use of expert testimony.

             An expert witness is helpful in many cases and is necessary in all scientific challenges.  Appropriate experts can help challenge field sobriety tests, breath tests, and blood tests.  

             Practitioners should make certain that their expert is appropriate for the defense.  It does not help to have a forensic toxicologist testify as to field sobriety tests, for example.

             While the procedure for qualifying an expert and eliciting expert opinion testimony is beyond the scope of this chapter, it is important to note that experts make their opinions based on hypothetical facts.  If the underlying facts are changed, the opinion should change.  Be careful not to forget that trial testimony can be different than expected.