The horizontal gaze nystagmus test (HGN) is often the first field sobriety test administered by the arresting officer.
If the police officer claims that you have “failed” the horizontal gaze nystagmus test then your attorney may attempt to raise a reasonable doubt about this result. The reasonable doubt may be based on any of several issues:
The horizontal gaze nystagmus test is a standardized field sobriety test. This means that the National Highway Traffic Safety Administration (NHTSA) has established that the tests must be given in a particular manner or they are not reliable. No freelancing is permitted.
Although the horizontal gaze nystagmus test is considered to be the most accurate of the various field sobriety tests, it is not perfect. Studies indicate that even when it is administered correctly its reliability rate is only 77%.
The following is a sample cross-examination of an officer about the horizontal gaze nystagmus test:
Q: Officer, are you familiar with a report by the Department of Transportation entitled “Improved Sobriety Testing”?
A: May I see it counselor?
Q: Please review the report and let me know whether you have seen it before.
A: Yes, I have seen this before; it was part of my training materials.
Q: That report contains an overview of the horizontal gaze nystagmus test as well as a test known as the walk-and-turn test and the one-leg stand test, correct?
Q: In this case you only administered the horizontal gaze nystagmus test to Mr. Jones, correct?
Q: You did not administer the walk-and-turn test, did you?
Q: You did not administer the one-leg stand test, did you?
Q: I would now like to read from the report which you have acknowledged you used in your training on HGN and ask if you agree with this statement: “All drivers whose behavior suggest impairment should be given at minimum, the gaze nystagmus test. The combination of all three tests (nystagmus, walk-and-turn and one-leg stand) . . . will give you firm data on which to base your arrest decision.” Do you agree with that statement?
Q: You did not perform this combination of the three tests, did you?
Q: Yet you arrested Mr. Jones after you administered the horizontal gaze nystagmus test, correct?
Q: The report goes on to indicate the appropriate testing procedures, does it not?
Q: Did you follow this testing procedure as outlined in the report when you administered the HGN test to Mr. Jones?
A: Basically, yes.
Q: Now you say “basically” you did. Did you, for example, score the subject on the point system used in the report?
A: I did not use that specific system, no.
Q: Officer, did you give the instructions to Mr. Jones as outlined in this report?
A: I do not recall if I gave those specific instructions.
Q: Do you agree, officer, that if the horizontal gaze nystagmus test is not performed properly as indicated in this report, the results are suspect?
A: I agree that the test should be conducted properly. I believe that I did conduct it properly.
Q: You do agree, however, that you deviated from the instructions set forth in this report, correct?
A: Yes, somewhat.
Q: Do you also agree, officer, that even if the HGN test is conducted properly according to this report it will only correctly classify about 77% of suspects as to whether they are drunk or sober?
A: I believe that’s correct.
Q: So in 23% of the cases it will not correctly classify suspects, correct?
Q: That means if there were one hundred people tested and the tests were performed exactly as set forth in this report, twenty-three of those one hundred people would be incorrectly classified as being either drunk or sober using the HGN test, correct?
A: I believe that’s correct.
Q: So 23%, or twenty-three people out of one hundred, would be improperly classified even if the test were performed exactly as set forth in this report, correct?
A: That’s correct.
This line of questioning has demonstrated that the officer did not conduct confirmatory tests that would have given him “firm data” for an arrest, and that even under ideal circumstances and properly conducted tests, the HGN test is only accurate 77% of the time.