What does it mean for something to be robust? Some of the synonyms of the word “robust” are strong, sturdy or powerful. In 2007 the National Highway Traffic Safety Administration (NHTSA) undertook a study to see just how “robust” or “powerful” the Horizontal Gaze Nystagmus Test was in the prediction of blood alcohol levels in humans. The study was titled “The Robustness of the Horizontal Gaze Nystagmus Test”, Southern California Research Institute, in conjunction with the U. S. Department of Transportation, authored by Marcelline Burns, Ph. D. They set out to find out, among other things, how varying certain elements of the test instructions would affect the interpretation of the testing. We will call it the “HGN Robustness Study” for brevity. The elements they varied were (1.) the closeness of the stimulus to the eyes, (2.) the elevation of the stimulus and (3.) the speed of the passing of the stimulus. Their conclusion was stated succinctly in their executive summary of the NHTSA study: “It is concluded that HGN is a robust phenomenon.” HGN Robustness Study, Introduction. As we will see this conclusion is at odds with the evidence.
For the study, the Southern California Institute took 108 subjects and brought them to different blood alcohol levels by varied doses of alcohol and orange juice. The dosing was “blind” as to the police officers; they did not observe the dosing. NHTSA chose experienced police officers who were SFST trained in the NHTSA curriculum and who had been so for a minimum of one year. Each was required to have administered SFSTs in the field at least 50 times. HGN Robustness Study, p.7. The officers would give the HGN to the subjects correctly, then vary an element in the test and keep a record of the scoring of that testing. Testing and recording were repeated for each of the varied elements. The data generated by the study showed a phenomenal number of false positive arrests. Instead of being robust, the data shows a robust odor and not a good one.
Of the 108 people tested, 76 of them had a BAC of less than .08%. NHTSA SFST criteria for arrest decision in the field is four or more clues on the HGN. NHTSA has previously stated that if an individual displays 4 or more clues on the HGN, it means that there is an 88% probability that the person’s BAC is .08% or greater. See, DWI Detection and Standardized Field Sobriety Testing, Participant Manual (NHTSA 2015 Edition), Session 8, p. 32. The “robust” study showed that, of the 76 people that tested below .08%, 51 of them were falsely arrested. That is to say, 67% of the people that were innocent (i. e. BAC below .08) were falsely arrested because they showed 4 clues or more on the HGN! This was when the test was done correctly! Thus there was a 67% false arrest rate when the stimulus was at the correct distance and speed – 12 to 15 inches from the subject’s nose, two inches above eye level, and a two-second pass to the nose when checking for smooth pursuit.
There were three groupings of people in the study with 36 people in each of the three groups. The number of people that were below .08% BAC and that scored four or more clues were as follows:
- 26 people under .08%, 20 with 4 or more clues = 77% false positive arrest
- 24 people under .08%, 13 with 4 or more clues = 54% false positive arrest
- 26 people under .08%, 18 with 4 or more clues = 69% false positive arrest
When the variations in the administration of the HGN were performed, the false arrest numbers went up even higher with the following results:
- Too fast (changing the speed of the stimulus from 2 seconds to 1 second) = 80% false arrest (It should be noted that they did not test by reduction of the speed of the stimulus.)
- Too high (4 inches instead of 2 inches above eye level) = 91% false arrest
- Too low (0 inches instead of 2 inches above eye level) = 77% false arrest
- Too close (10 inches instead of 12 to 15 inches from face) = 88% false arrest
- Too far (20 inches instead of 12 to 15 inches from face) = 96% false arrest
The BAC ranges were varied with some as low as .016% and .019%. In fairness, I should point out that some of the false positives were as high as .079%. Nevertheless, the NHTSA manual sticks with the proposition that 4 or more clues on the HGN means that there is an 88% probability that your BAC is .08% or higher. Interestingly, the 88% figure represents an increase from 77% from the 2006 NHTSA SFST Student Manual, presumably the change was made due to the San Diego Validation Study. See, Burns, M., Ph.D. and J. Stutter, Validation of the Standardized Field Sobriety Test Battery at BACs below 0.10%, August 1998.
With data this disastrous for NHTSA, what do you think they did in their reporting of the outcome of the study? They changed the criteria for what constitutes a false positive arrest! Instead of 4 clues and a .08 standard, the subject had to be below a .06% BAC and show 6 clues! This is absurd when the arrest criteria used by the street cops across our country are trained that 4 or more clues mean that a person has an 88% probability of having a BAC of .08% or higher. The only thing robust here is NHTSA’s blind commitment to the reliability of HGN no matter the evidence.
About the Author: Phillip B. Price, Sr. the senior partner at Price, Flowers & Ward Law Firm. He is a former Dean of the National College for DUI Defense (NCDD), the premier national organization for DUI defense attorneys. He is the only attorney in North Alabama who is Board Certified as a Specialist in DUI Defense. He has been representing citizens accused of DUI in Alabama for over 35 years. In 2012, he received the prestigious Erwin-Taylor Award from the NCDD. He is the author the Alabama DUI Handbook, published by Thomson Reuters ®.
If you would like to contact the author, please visit: http://www.alabamaduiattorney.com/