Field sobriety exercises are scored by counting clues, and the way the counting works is itself a weakness. Horizontal gaze nystagmus has six possible clues, the walk and turn has eight, and the one leg stand has four. The training treats four of six on the eye test, two of eight on the walk and turn, and two of four on the one leg stand as the points where an officer decides a person is likely impaired. Under State v. Meador, 674 So. 2d 826 (Fla. 4th DCA 1996), none of this is a scientific score, only the officer’s lay observation.
Two things stand out. The thresholds are low, and the clues are binary. A clue is a checkbox the officer either marks or does not, not a measurement of how much someone is affected. Because the lines sit at two clues for the balance exercises, a single debatable checkbox can move a person from below the line to above it.
The thresholds are low and the clues are binary. A single extra checkbox can move you across the line, which is why how the clues were counted matters so much.
The stage error on the walk and turn
The eight walk-and-turn clues are split between the instruction stage, where you stand heel to toe and listen, and the walking stage, where you step, turn, and step back, and the two stages are scored separately. A common mistake is recording a clue in the wrong stage or counting the same behavior twice, which quietly inflates the total. Matched against the trained method, those miscounts can drop a result back below the decision point.
The pass-fail myth
Officers and reports often slip into the language of passing, failing, and points. Meador forbids exactly that, because calling these exercises tests that a person passes or fails borrows a scientific certainty they do not have. Holding the officer to observation language, rather than the vocabulary of a graded exam, changes how a jury hears the whole thing.
What a single clue is really worth
Two clues out of eight ends the walk and turn, and two out of four ends the one leg stand, which sounds strict until you ask how often a sober person hits those marks. In the government’s own San Diego study, among the drivers who were under 0.08, more than half showed two or more clues on the walk and turn and four in ten showed two or more on the one leg stand. So when an officer testifies that you had “two clues,” what a jury needs to hear is that a sober person clears that same bar about as easily as you did. The count sounds like arithmetic, and it feels like a grade, but it is a set of small judgment calls that lands on the wrong side of the line for ordinary people all the time.
Clues on exercises that have none
On the non-standardized exercises, the finger to nose, the modified Romberg, the finger count, and the alphabet, there are no official clues and no scoring at all. A clue there is entirely the officer’s impression. I keep that distinction sharp for the jury, because a checkbox on an unvalidated exercise is not evidence of a blood alcohol level.
The Arithmetic Behind the Clues
Each exercise carries a fixed clue count and a fixed threshold: four or more of six clues on HGN, two or more of eight on the walk and turn, two or more of four on the one leg stand. NHTSA’s research ties those thresholds to classification rates of about 77, 68, and 65 percent respectively, all measured against the old 0.10 standard rather than today’s 0.08, and all validated to predict a number, not to measure impairment.
The flattering percentages also hide the direction of the errors. Calculated from the 1998 San Diego study’s own tables, the false positive rates for drivers who were under 0.08 were 37 percent on HGN, 52.6 percent on the walk and turn, and 41.3 percent on the one leg stand. The study never published those numbers, and the study itself was never peer reviewed.
Notice what the thresholds mean in practice. On the one leg stand, half of the available clues puts you in the arrest column. On the walk and turn, a quarter of them does. The scoring is binary, a clue either counts or it does not, and there is no column on the score sheet for wind, headlights, footwear, or fear. That is not cross-examination spin; it is how the scoring instrument is built, by the agency that built it.
When I cross the officer on the count, I take it apart one clue at a time rather than argue about the total, with you sitting right there so the jury sees a person and not a score sheet. Under State v. Meador, 674 So. 2d 826, 833 (Fla. 4th DCA 1996), no one in that room may tell them you “passed” or “failed” or racked up “points,” because those words give a roadside guess “an aura of scientific validity” it has not earned. I teach how these clues are counted, so I can show the jury that two out of eight comes down to a coin toss the State is asking them to trust with your life.
Questions About Clues and Scoring
How are field sobriety tests scored?
By counting clues. Horizontal gaze nystagmus has six possible clues, the walk and turn has eight, and the one leg stand has four. The training treats four of six, two of eight, and two of four as the decision points for likely impairment.
Is a clue the same as a measurement?
No. A clue is a yes-or-no checkbox the officer marks, not a measured value. Counting two clues does not measure how impaired someone is, and the thresholds are low enough that one debatable checkbox can cross the line.
What is the instruction-stage error on the walk and turn?
The eight clues are split between a standing instruction stage and a walking stage, scored separately. Officers sometimes record a clue in the wrong stage or double-count, which inflates the total against the trained method.
Should an officer say I passed or failed?
No. Under State v. Meador an officer should not tell the jury you passed, failed, or scored points, because that lends a scientific certainty the exercises do not have. They are observations, not graded tests.
What is a clue on a non-standardized test worth?
Nothing validated. Exercises like the finger to nose or the alphabet have no official clues or scoring at all, so a clue there is purely the officer's impression.
How do you challenge the scoring?
I check each clue against the trained definition, the correct stage, and the video, and I hold the officer to the rule that these are observations rather than a pass-fail score.
Related: the walk and turn, administration errors, the non-standardized exercises, HGN as scientific evidence, and why the results are unreliable.
Is two clues really enough to fail?
That is the number the training uses, but it means less than it sounds. In the San Diego study, more than half of the sober drivers under 0.08 showed two or more walk and turn clues, so a passing number of clues is common in people who had nothing to drink.
This page is general information, not legal advice, and it does not create an attorney-client relationship. Field sobriety exercises in Florida are governed by case law and by section 316.193, Florida Statutes. Procedures and rules change, and every case turns on its own facts. Past results do not guarantee a similar outcome.
