The Walk-and-Turn Test and How It Is Scored
The walk-and-turn is a divided attention test: nine heel-to-toe steps along a line, a prescribed turn on one foot, and nine steps back, while the officer counts eight defined clues. Two clues out of eight is scored as a failure, and two of them can be recorded during the instruction stage before you take a single step. The sharpest challenge is not statistical. Three of the eight clues depend on a line actually existing, and the protocol calls for a reasonably dry, hard, level, non-slippery surface. A crowned highway shoulder, gravel, or wet pavement is not that, and difficulty caused by the ground is not evidence about the driver.
You can fail the walk and turn before you take a step.
Two of the eight clues are scored during the instructions, while you are standing still in an awkward heel-to-toe stance listening to a stranger talk. Two clues is a failing score. So a person who breaks that stance and starts walking a moment early has failed the test without walking a single line.
The instructions themselves contain fourteen separate things to remember, delivered once, at the roadside, at night. It is called a divided attention test, and it divides attention exactly as designed. Whether that measures alcohol is a different question.
How the Test Works, Stage by Stage
The walk-and-turn is built to divide your attention. It asks you to hold a balance position while processing and remembering a set of verbal instructions, then to walk while counting out loud and keeping track of how many steps you have taken. The mental side is comprehension, processing, and memory. The physical side is balance and coordination. Doing either alone is easy. Doing both at once is the test.
It runs in two stages, and both are scored.
The instruction stage. You put your left foot on the line, then your right foot on the line ahead of it, with the heel of the right foot against the toe of the left. Arms down at your sides. You hold that position while the officer finishes the instructions, and you do not start walking until told to. The officer is supposed to ask whether you understand so far, and to demonstrate.
The walking stage. Nine heel-to-toe steps down the line. Then a turn: keep the front foot planted on the line and turn by taking a series of small steps with the other foot. Then nine heel-to-toe steps back. While you walk, arms stay at your sides, you watch your feet the whole time, and you count each step out loud. Once you start, you do not stop until you are done.
The officer has obligations too. Give the full instructions. Make certain they are understood. Demonstrate the exercise. Observe from a safe distance and limit his or her own movement, because the officer moving around can itself distract the person being tested.
The test also has physical requirements. It needs a designated straight line, real or imaginary. It needs a reasonably dry, hard, level, non-slippery surface. And it needs enough room to take nine heel-to-toe steps in each direction.
The Eight Clues and the Two-Clue Failure Line
Count what you were actually asked to do and you get fourteen separate things to hold in mind at once: stand on a line, right foot in front of left, heel to toe, arms at your sides, do not begin until told, take nine steps, touch heel to toe, return nine steps, leave your front foot on the line at the turn, take a series of small steps, do not stop once you begin, watch your feet at all times, count each step out loud, and start your count from the heel-to-toe position.
You are scored on eight clues. Each has a definition, and the definitions are narrower than officers usually apply them:
- Cannot keep balance during the instructions. Your feet must actually break apart. Swaying, or using your arms to steady yourself, is not this clue if you hold the heel-to-toe position.
- Starts too soon. This counts only if the officer actually told you not to begin until instructed. Officers often leave that out.
- Stops while walking. Requires pausing for several seconds. Walking slowly is not a clue.
- Does not touch heel to toe. Requires a gap of more than half an inch between heel and toe.
- Steps off the line. One foot has to come entirely off the line.
- Uses arms for balance. Raising one or both arms more than six inches from your sides.
- Improper turn. Actually removing the front foot from the line while turning, or spinning or pivoting instead of taking a series of small steps.
- Incorrect number of steps. Taking more or fewer than nine steps in either direction.
Two of those eight is a failure. And two of the eight, cannot keep balance and starts too soon, are only scoreable during the instruction stage. Which means it is possible to fail this test before you have taken a single step, by shifting out of an awkward stance and starting early after hearing the instructions once.
Each clue counts only once no matter how many times it happens, which cuts slightly the other way and is worth knowing.
Then consider what the stance itself is doing. Standing with your feet together gives you a narrower base than you would ever naturally choose, and standing heel-to-toe is narrower still. Holding your arms pinned at your sides removes the thing every human body reaches for when balance goes. Nobody walks a balance beam with their arms down. Two of the three standardized tests are, by design, exercises in making balance harder.
The Line That Was Never There
Here is the sharpest single point on this page, and it is not an argument about statistics.
This is a walk-the-line test. Three of the eight clues depend on a line existing: steps off the line, improper turn, and the whole heel-to-toe geometry. The protocol requires a designated straight line, and it can be real or imaginary. But it has to be designated. The officer has to establish it and tell you where it is.
Officers routinely do not. They give the instructions, skip any mention of a line, watch the person walk across an unmarked stretch of asphalt, and then write down that the person stepped off the line.
You cannot step off a line that does not exist.
That is not a technicality about wording. It is the difference between a scored clue and an officer’s after-the-fact impression of where a line would have been if there had been one. And it is provable, because the instructions the officer gave are on the video.
The same logic runs through the rest of the administration. The clue “starts too soon” requires that the officer told you not to start. The clue “incorrect number of steps” depends on the officer telling you to count your first step from the heel-to-toe position as one, because without that instruction people count the planted foot as one, end up a step off, and turn in the wrong place. Leave the instruction out and you have manufactured the clue.
There is one more thing worth watching for on the footage, and it lands with juries. The officer is required to demonstrate this test. Officers demonstrating it have been known to break the heel-to-toe stance and raise their arms for balance while doing so, which is to say that by NHTSA’s own scoring, the officer produced two clues and failed the test he was about to give you.
None of this means the test is meaningless. It means the accuracy claims attached to it hold only when the test is administered in the prescribed manner, with the standardized clues, applied by the standardized criteria. NHTSA’s own materials say exactly that: change any one of those elements and the validity is compromised. The people in the Colorado, Florida, and San Diego validation studies were screened candidates, tested properly, and given a line. In the Florida study, 70% of subjects who were under 0.08 still produced two or more clues on this test.
Surface, Footwear, and Physical Condition
The walk-and-turn was validated on a reasonably dry, hard, level, non-slippery surface. A highway shoulder with a drainage crown is not that. Neither is gravel, wet pavement, grass, or cracked asphalt. Difficulty that comes from the ground is not evidence about the driver, but the scoring sheet cannot tell the two apart.
Footwear matters for the same mechanical reason, and the protocol acknowledges it: anyone wearing heels more than two inches high is supposed to be given the chance to take their shoes off. Work boots, wedges, and sandals are not covered by that instruction and create their own problems on a heel-to-toe line.
Physical condition matters too. The original research, Development and Field Test of Psychophysical Tests for DWI Arrests, indicated that people over 65, and people with back, leg, or middle ear problems, had difficulty with this test. Fewer than 1.5% of the subjects in that research were over 65. An older edition of the training manual listed wind and weather, the subject’s age or weight, and the subject’s footwear as conditions that may interfere with performance. Later editions replaced that list with a statement that field validation studies have shown environmental conditions do not affect a subject’s ability to perform the test, without identifying which studies, and no new validation studies have been done.
The full factor list, one at a time, with what the source says about each, belongs on Who Fails Field Sobriety Tests While Sober.
Medical Conditions That Defeat the Test
An officer is trained to ask, before administering the balance tests, whether you have any physical problem or disability that would affect your ability to perform, and the training names back, leg, and inner ear problems specifically. If the answer is yes, the officer has options: alternative tasks, or simply not scoring you on a test you were never a suitable candidate for.
In practice the question is often asked as a formality, asked too fast to answer, or not asked at all. Worse, drivers who do disclose a bad knee or a balance disorder are frequently run through the test anyway and scored as though they had said nothing.
A diagnosed condition is a stronger and different argument than an ordinary characteristic, because it is provable with medical records rather than inferred from footage. Vestibular disorders, neurological conditions, orthopedic injuries, and the medications that treat them are covered on Medical Conditions Defense.
The Other Two Standardized Tests
The walk-and-turn is one of three. The one-leg stand is the other balance test, scored on four clues with the same two-clue failing line, and it has its own set of instructions officers commonly add to: The One-Leg Stand Test. The eye test is different in kind, treated by Texas courts as scientific evidence with a legal limit on what it can prove: The HGN Eye Test. How the three fit together, where the battery came from, and whether you had to perform any of them are on Field Sobriety Tests.
Key Terms
The words that come up most on this page, in plain English.
- Walk and turn test
- A divided attention field sobriety test requiring nine heel-to-toe steps along a line, a turn on one foot, and nine steps back. It has an instruction stage and a walking stage, eight standardized clues, and a failing threshold of two clues.
- Divided attention test
- A test designed to require a subject to handle mental and physical tasks simultaneously. On the walk and turn the mental tasks are comprehension, information processing, and recall, and the physical tasks are balance and coordination while standing, walking, and turning.
- Instruction stage
- The first stage of the walk and turn, in which the subject stands in a heel-to-toe position with arms at their sides while instructions are given. Two of the eight clues, losing balance and starting too soon, are scored during this stage only.
- Clue threshold
- The precise definition each walk and turn clue must meet to be scored, such as one foot entirely off the line, arms raised more than six inches, a heel-to-toe gap greater than half an inch, or a pause of several seconds.
The other guides in this section.
The Breath Test
The Blood Test
The Roadside Tests
Drug Cases
The Science of the Number
The science on this page is the raw material. The moves that turn it into a suppressed result, an excluded expert or a dismissal live in Defenses.
Charged With a DWI in Texas?
We're available 24/7 across all six Texas offices.
Request a Free Case Evaluation
Tell us what happened. We'll respond as soon as possible.