Just arrested for DUI? In Arizona you have 30 days to request your MVD hearing. Start here →
DUIINFO

Roadside Evidence, Examined

Field Sobriety Tests

Subjective tests, scored by the person arresting you.

Only three roadside tests are standardized by NHTSA — and even under laboratory conditions, none of them is more than 77% accurate. Learn how each test is supposed to be administered, what officers score as “clues,” and the long list of factors that can make a sober person fail.

3
NHTSA-Standardized Tests
65–77%
Accuracy Even in the Lab
Voluntary
FSTs in Arizona — No Refusal Penalty

The Three Standardized Tests

NHTSA’s Standardized Field Sobriety Test (SFST) battery contains exactly three tests. Each has a prescribed administration procedure, a fixed list of scorable “clues,” and a decision point. Deviate from the procedure, and the published accuracy rates no longer apply — NHTSA’s training materials say so themselves.

Horizontal Gaze Nystagmus (HGN)

~77% lab accuracy6 clues (3 per eye)

How It’s Administered

The officer holds a stimulus — usually a pen or fingertip — about 12 to 15 inches from your face, slightly above eye level, and moves it slowly side to side while watching your eyes. You are told to follow the stimulus with your eyes only, without moving your head.

What Officers Score as Clues

  • Lack of smooth pursuit — the eye jerks as it tracks the stimulus instead of following it smoothly
  • Distinct and sustained nystagmus at maximum deviation — the eye jerks noticeably when held at the far corner for at least four seconds
  • Onset of nystagmus prior to 45 degrees — jerking begins before the eye reaches a 45-degree angle

Decision point: 4+ clues suggests BAC ≥ 0.08%

Where It Falls Apart

  • Dozens of medical conditions cause nystagmus with zero alcohol: inner-ear problems, brain injury, hypertension, caffeine, nicotine, and certain prescription medications
  • Flashing police lights or passing traffic in your field of view can trigger optokinetic nystagmus — NHTSA's own manual says the test should face you away from them
  • Timing errors are common: each pass has a prescribed speed (roughly two seconds out, two seconds back) and the officer must hold the stimulus at maximum deviation for a minimum of four seconds
  • Holding the stimulus too close, too high, or moving it too fast invalidates the clue

Walk-and-Turn (WAT)

~68% lab accuracy8 clues

How It’s Administered

A divided-attention test in two stages. First, the instruction stage: you stand heel-to-toe with arms at your sides while the officer explains and demonstrates. Then the walking stage: nine heel-to-toe steps down a real or imaginary line, a prescribed series of small pivot steps to turn, and nine heel-to-toe steps back — counting out loud, watching your feet, arms at your sides.

What Officers Score as Clues

  • Cannot keep balance during instructions
  • Starts before instructions are finished
  • Stops while walking
  • Misses heel-to-toe (more than half an inch)
  • Steps off the line
  • Uses arms for balance (raises more than six inches)
  • Makes an improper turn
  • Takes the wrong number of steps

Decision point: 2+ clues suggests BAC ≥ 0.08%

Where It Falls Apart

  • NHTSA's validation studies excluded people over 65, those more than 50 pounds overweight, and those with back, leg, or inner-ear problems — yet officers give the test to everyone
  • The standardized conditions require a hard, dry, level, non-slippery surface and a visible line — a sloped freeway shoulder at 1 a.m. is none of those things
  • Heels higher than two inches: NHTSA says you should be offered the chance to remove them. Many officers never do
  • An officer who demonstrates the turn incorrectly, or never says 'keep your arms at your sides,' has scored you on instructions you were never given

One-Leg Stand (OLS)

~65% lab accuracy4 clues

How It’s Administered

You stand with feet together and arms at your sides, then raise one foot approximately six inches off the ground, toe pointed, both legs straight, and count out loud — 'one thousand one, one thousand two…' — while looking at your raised foot. The officer times the test for 30 seconds.

What Officers Score as Clues

  • Sways while balancing
  • Uses arms for balance (raises more than six inches)
  • Hops to maintain balance
  • Puts the raised foot down before time expires

Decision point: 2+ clues suggests BAC ≥ 0.08%

Where It Falls Apart

  • Standing on one leg for 30 seconds is hard for many sober people — the same age, weight, and injury exclusions from the WAT studies apply here and are routinely ignored
  • If the officer fails to time the full 30 seconds, the scoring is invalid — most impairment-related foot-downs occur after the 25-second mark
  • 'Sways while balancing' has no objective measurement; it is whatever the officer says it is
  • Gravel, rain, wind, cold, and darkness all degrade balance and were never part of the validation conditions

What Invalidates a Field Sobriety Test

The accuracy figures above were measured under controlled conditions on a screened population. The roadside is neither. Every factor below is a documented source of false “clues” — and a documented basis for challenging the test in court.

Medical Conditions

Inner-ear disorders, vertigo, neuropathy, diabetes, prior head injuries, knee and back problems, and many medications independently cause the exact 'clues' officers score as impairment.

Fatigue

Most DUI stops happen late at night. Sleep deprivation alone produces nystagmus, sway, and balance loss — the same signs the tests attribute to alcohol.

Footwear & Surface

Heels, boots, sandals, gravel, slopes, and painted (slick) lines all violate the standardized test conditions NHTSA requires for valid scoring.

Weather & Environment

Wind, rain, cold, uneven shoulders, and strobing emergency lights degrade performance and can directly trigger false HGN clues.

Age & Weight

NHTSA's own validation research excluded subjects over 65 and those 50+ pounds overweight. If that describes you, the published accuracy rates were never measured on people like you.

Officer Instruction Errors

The tests are only validated when administered in the standardized manner. Skipped demonstrations, wrong timing, and improvised scoring break the protocol — and the value of the evidence.

Non-Standardized Tests Carry Even Less Weight

Anything outside the three-test SFST battery — alphabet recitation, finger-to-nose, counting exercises — has no NHTSA validation, no standardized scoring, and no published accuracy rate. When an officer relies on these, there is no science behind the opinion at all, only the officer’s say-so. Courts and juries can be shown exactly that.

Not validated

Alphabet Test

Reciting the alphabet from a given letter, or backwards. Never validated by NHTSA — many sober people cannot say the alphabet backwards on demand.

Not validated

Finger-to-Nose

Eyes closed, head tilted back, touching your nose on command. No standardized scoring criteria exist; the 'result' is pure officer opinion.

Not validated

Counting / Finger Count

Counting backwards or touching fingers to thumb in sequence. Divided-attention in theory, but with no validated clue system behind it.

Not validated

Romberg Balance

Standing with eyes closed, head back, estimating 30 seconds. Borrowed from a 19th-century neurological exam — not one of the three standardized tests.

Know Your Rights

You Can Refuse FSTs in Arizona

Field sobriety tests are voluntary. Arizona law attaches no statutory penalty — no license suspension, no separate offense — to politely declining them. Officers are not required to tell you this, and most don’t.

Chemical tests are different. Under Arizona’s implied consent law, refusing a breath or blood test after arrest triggers a one-year license suspension — and you have just 30 days to request the MVD hearing that contests it (A.R.S. § 28-1385). Never confuse the two categories: declining roadside balance tests costs you nothing by statute; refusing the chemical test costs you your license for a year.

What to do in the first 24 hours

Arizona Case Law

HGN Needs a Foundation

In State ex rel. Hamilton v. City Court of Mesa, the Arizona Supreme Court set the rules for HGN evidence: it may be used to support probable cause for arrest and to corroborate impairment, but only when the State lays a proper foundation — an officer qualified in the technique who administered it in the standardized manner. HGN cannot be offered to prove a specific numeric BAC.

That foundation requirement is a pressure point. Training records, certification lapses, and body-cam footage showing a rushed or improvised test all open the door to excluding the HGN evidence entirely.

More defense strategies

From Roadside to Courtroom

How FST Errors Become Motions

Every administration error documented above maps to a specific legal filing. A test given off-protocol can support a motion to suppress the FST results, or — if the stop itself rested on the tests — a challenge to probable cause for the arrest. Foundation gaps in the HGN feed a motion in limine to keep the jury from ever hearing about it. And when FST testimony does come in, properly framed jury instructions remind jurors that these are opinion-based screening tools, not measurements.

Suppression Motions

Attack the stop, the detention, and FST results obtained through non-standardized administration or without reasonable suspicion.

Motions in Limine

Keep unreliable test evidence — non-standardized tests, unqualified HGN testimony, BAC-correlation opinions — away from the jury before trial starts.

Jury Instructions

Proposed instructions that frame FSTs and HGN for what they are: subjective screening tools with documented error rates.

Turn Test Errors Into Filings

The Motion Bank’s trial kit includes FST and HGN jury instructions and a 10-motion limine set built around exactly the failure modes on this page — administration errors, foundation gaps, and non-standardized tests.

Explore the Motion Bank

Field Sobriety Test FAQ

Can I refuse field sobriety tests in Arizona?

Yes. Field sobriety tests are voluntary in Arizona, and there is no statutory penalty — no license suspension, no separate charge — for politely declining them. That is fundamentally different from chemical tests (breath or blood), where refusal triggers a one-year implied-consent suspension under A.R.S. § 28-1385. Officers are not required to tell you the FSTs are optional, and most never do.

How accurate are field sobriety tests?

Under controlled laboratory conditions, NHTSA's own research puts HGN at about 77% accurate, the Walk-and-Turn at about 68%, and the One-Leg Stand at about 65% at predicting a BAC of 0.08% or more. That means even when administered perfectly, the Walk-and-Turn is wrong roughly one time in three — and roadside conditions are never laboratory conditions.

What medical conditions can cause a false 'failure'?

Inner-ear disorders, vertigo, prior head injuries, hypertension, diabetes, neuropathy, arthritis, and knee, hip, or back problems can all produce the sway, balance loss, and eye movements officers score as clues. Dozens of conditions and common medications cause nystagmus with no alcohol involved at all. NHTSA's validation studies excluded many of these populations entirely.

Is HGN admissible in court?

It depends on the state and the foundation laid. In Arizona, State ex rel. Hamilton v. City Court of Mesa allows HGN evidence to support probable cause and to corroborate impairment, but only with a proper foundation — a qualified officer who administered the test in the standardized manner. HGN generally cannot be used to establish a specific numeric BAC. Foundation gaps are a classic target for a motion in limine.

The officer never told me the tests were voluntary. Does that matter?

It can. While officers generally are not obligated to announce that FSTs are optional, how the tests were presented bears on whether your participation was truly voluntary and on the weight of the evidence. Coercive framing, combined with administration errors, strengthens motions to suppress or limit FST testimony.

I 'failed' the field tests. Is my case hopeless?

No. FST results are opinion evidence layered on subjective scoring, and they are attacked successfully every day: wrong instructions, improper conditions, excluded populations, medical explanations, and video that contradicts the report. A 'failure' on paper often looks very different frame-by-frame on the body-cam footage.