Field Sobriety Tests Do Not Accurately Determine Whether Driving Ability is Impaired.
Everybody is different. NHTSA developed these tests in a controlled environment. Their “test subjects” were dosed with alcohol and NHTSA studied the test subjects' ability to perform these tasks at various blood alcohol levels. The conditions to which NHTSA's test subjects were subjected are far different from the testing conditions to which a DUI suspect is subjected.
First, NHTSA's initial test subjects were not under investigation for DUI. They were volunteers who agreed to drink alcohol and perform various tasks in a gymnasium. A typical DUI suspect has been pulled over by the police. He begins his experience with a sinking feeling as he sees the flashing blue lights behind him. He pulls his car to the side of the road and the officer tells him to step out of his car to perform field sobriety tests. NHTSA's test subjects did not experience a DUI suspect's level of anxiety. NHTSA's volunteers did not begin their experience in by being ordered out of their car. Their heart did not quicken as they saw the blue lights behind them. Their hands did not begin to shake as the officer told them to get out of the car. They did not have an officer with a gun and a nightstick ready to take them to jail for failing to satisfactorily perform the tests. Most jurors understand that Field Sobriety tests are difficult under the best of circumstances. Jurors also understand that your performance on field sobriety may be the result of nervousness rather than alcohol impairment.
Second, FST's do not test your driving ability. You are not guilty of DUI unless the State can prove that you were incapable of driving safely (or you were above the legal limit). Officers say that poor performance on FST's is an indicator of impaired driving ability. This is simply not true. People practice driving everyday. DUI suspects only get one chance to perform FST's. Officers do not let DUI subjects practice these tests. Even worse, the officers never tell the suspect what is required to pass the test so the suspect never knows what is expected of him. Again, everybody is different. Most people perform less well on FST's under such stressful conditions. Therefore, there is no direct correlation between driving ability and performance on FST's. One does not stand on one leg for thirty seconds while they drive a car. These are two completely unrelated skills.
NHTSA claims to have conducted “field” validation studies under arrest conditions; however their findings are unscientific as shown by their own data. Police officers often claim FST's are 91-94% accurate in making the arrest decision at a .08 level. To support this claim, they cite the “California Study”. However, researchers used faulty science to “validate” FST's. Most importantly they validated FST's at a .08 level by using group of subjects with a sample mean (average) of .150 grams of alcohol. (see NHTSA's table below). To validate FST's at the .08 level, the researchers should have studied subjects closer to .08. Instead the researchers studied subjects who were obviously impaired.
To provide a real-world example of NHTSA's flawed method, suppose that you want to know whether NASCAR fans can accurately predict whether cars exceed a 65-mile per hour limit. To scientifically answer that question, your NASCAR fans must see cars traveling near 65 miles per hour. The average speed of the cars tested should be close to 65 mph, the desired discrimination point, and the fans must guess whether each car exceeded 65 mph. A validation study is meaningless if the NASCAR fans watch cars traveling at 135 mph. 100% of the NASCAR fans will accurately predict the cars exceed 65 mph. The same is true when the slower cars travel at 20 mph. The fans will correctly predict those cars are slower than 65. Although the fans would accurately predict speed over and under 65 mph under those circumstances, the study does nothing to truly validate predictions at 65 mph.
Similarly, NHTSA studied DUI arrestees with mean BAC's of .150 and non-arrest subjects with a mean BAC of .045 in the California study. In other words, those who were arrested, were obviously over a .08 since the average arrestee was .150 while the average non-arrestee was .045.
This site contains additional information on DUI suspensions, DUI penalties, sobriety tests, Georgia DUI statues and how they are applied in Gwinnett County, Atlanta, Fulton County, Barrow County, Hall County.