STANDARDIZED FIELD SOBRIETY TESTING

ONE-DAY REFRESHER

8:30a.m. – 8:45a.m. Welcome, Introductions, Purpose Statement

8:45a.m. – 10:15a.m. Phase 1 – Vehicle in Motion

Phase 2 – Personal Contact

10:15a.m. – 10:30a.m. Break

10:30a.m. – 12:00p.m. Concepts & Principles of SFSTs

Phase 3 – Pre-Arrest Screening

12:00p.m. – 1:00p.m. Lunch

1:00p.m. – 1:30p.m. Demonstration & Dry Run

1:30p.m. – 2:30p.m. Dry Lab Workshop

2:30p.m. – 2:45p.m. Break

2:45p.m. – 3:30p.m. Dry Lab Workshop

3:30p.m. – 4:30p.m. Exam – Practical & Written

This program was created and sponsored by the Illinois Department of Transportation -
Division of Traffic Safety,
Traffic Safety Resource Prosecutor Program
 PREFACE

The procedures outlined in this manual describe how the Standardized Field
Sobriety Tests (SFSTs) are to be administered under ideal conditions. We recognize
that the SFSTs will not always be administered under ideal conditions in the field,
because such conditions will not always exist. Even when administered under less
than ideal conditions, they will generally serve as valid and useful indicators of
impairment. Slight variations from the ideal, i.e., the inability to find a perfectly
smooth surface at roadside, may have some affect on the evidentiary weight given to
the results. However, this does not necessarily make the SFSTs invalid.
INTRODUCTION AND OVERVIEW
INTRODUCTION AND OVERVIEW

Upon successfully completing this session, the participant will be able to:

o State the goals and objectives of the course.

o Describe the course schedule and activities.
o Demonstrate their pre-training knowledge of course topics.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Welcoming Remarks and Objectives o Instructor-Led Presentations

B. Administrative Details

2
 DWI DETECTION AND STANDARDIZED FIELD SOBRIETY TESTING

TRAINING GOALS AND OBJECTIVES

1. Ultimate Goal

To increase deterrence of DWI violations, and thereby reduce the number of

crashes, deaths and injuries caused by impaired drivers.

2. Enforcement-Related Goals

a. Understand enforcement's role in general DWI deterrence.

b. Understand detection phases, clues and techniques.
c. Understand requirements for organizing and presenting testimonial and
documentary evidence in DWI cases.

3. Job Performance Objectives

As a result of this training, participants will become significantly better able to:

a. Recognize and interpret evidence of DWI violations.

b. Administer and interpret Standardized Field Sobriety Tests.
c. Describe DWI evidence clearly and convincingly in written reports and verbal
testimony.

4. Enabling Objectives

In pursuit of the job performance objectives, participants will come to:

a. Understand the tasks and decisions of DWI detection.

b. Know and recognize typical vehicle maneuvers and human indicators
symptomatic of DWI that are associated with initial observation of vehicles in
operation.
c. Know and recognize typical reinforcing maneuvers and indicators that come
to light during the stopping sequence.
d. Know and recognize typical sensory and other clues of alcohol and/or other
drug impairment that may be seen during face-to-face contact with DWI
suspects.
e. Know and recognize typical behavioral clues of alcohol and/or other drug
impairment that may be seen during the suspect's exit from the vehicle.
f. Understand the role and relevance of psychophysical testing in pre-arrest
screening of DWI suspects.

3
g. Know and carry out appropriate administrative procedures for validated
divided attention psychophysical tests.
h. Know and carry out appropriate administrative procedures for the Horizontal
Gaze Nystagmus test.
i. Know and recognize typical clues of alcohol and/or other drug impairment
that may be seen during administration of the Standardized Field Sobriety
Tests.

4
 ATTACHMENT

GLOSSARY OF TERMS

ALVEOLAR BREATH - Breath from the deepest part of the lung.

BLOOD ALCOHOL CONCENTRATION (BAC) - The percentage of alcohol in a person's

blood.

BREATH ALCOHOL CONCENTRATION (BrAC) - The percentage of alcohol in a person’s

breath, taken from deep in the lungs.

CLUE - Something that leads to the solution of a problem.

CUE - A reminder or prompting as a signal to do something. A suggestion or a hint.

DIVIDED ATTENTION TEST - A test which requires the subject to concentrate on both
mental and physical tasks at the same time.

DWI/DUI -The acronym "DWI" means driving while impaired and is synonymous
with the acronym "DUI", driving under the influence or other acronyms used to
denote impaired driving. These terms refer to any and all offenses involving the
operation of vehicles by persons under the influence of alcohol and/or other drugs.

DWI DETECTION PROCESS - The entire process of identifying and gathering evidence to
determine whether or not a suspect should be arrested for a DWI violation. The DWI
detection process has three phases:

Phase One - Vehicle In Motion

Phase Two - Personal Contact
Phase Three - Pre-arrest Screening

EVIDENCE - Any means by which some alleged fact that has been submitted to
investigation may either be established or disproved. Evidence of a DWI violation may be
of various types:

a. Physical (or real) evidence: something tangible, visible, or audible.

b. Well established facts (judicial notice).
c. Demonstrative evidence: demonstrations performed in the courtroom.
d. Written matter or documentation.
e. Testimony.

FIELD SOBRIETY TEST - Any one of several roadside tests that can be used to determine
whether a suspect is impaired.

HORIZONTAL GAZE NYSTAGMUS (HGN) - An involuntary jerking of the eyes

5
as they gaze toward the side.

ILLEGAL PER SE - Unlawful in and of itself. Used to describe a law which makes
it illegal to drive while having a statutorily prohibited Blood Alcohol Concentration.

NYSTAGMUS - An involuntary jerking of the eyes.

ONE-LEG STAND (OLS) - A divided attention field sobriety test.

PERSONAL CONTACT - The second phase in the DWI detection process. In this
phase the officer observes and interviews the driver face to face; determines
whether to ask the driver to step from the vehicle; and observes the driver's exit and
walk from the vehicle.

PRE-ARREST SCREENING - The third phase in the DWI detection process. In

this phase the officer administers field sobriety tests to determine whether there is
probable cause to arrest the driver for DWI, and administers or arranges for a
preliminary breath test.

PRELIMINARY BREATH TEST (PBT) - A pre-arrest breath test administered

during investigation of a possible DWI violator to obtain an indication of the
person's blood alcohol concentration.

PSYCHOPHYSICAL - "Mind/Body." Used to describe field sobriety tests that

measure a person's ability to perform both mental and physical tasks.

STANDARDIZED FIELD SOBRIETY TEST BATTERY - A battery of tests,

Horizontal Gaze Nystagmus, Walk-and-Turn, and One-Leg Stand, administered
and evaluated in a standardized manner to obtain validated indicators of
impairment based on NHTSA research.

TIDAL BREATH - Breath from the upper part of the lungs and mouth.

VEHICLE IN MOTION - The first phase in the DWI detection process. In this
phase the officer observes the vehicle in operation, determines whether to stop
the vehicle, and observes the stopping sequence.

VERTICAL GAZE NYSTAGMUS - An involuntary jerking of the eyes ( up and

down) which occurs when the eyes gaze upward at maximum elevation.

WALK-AND-TURN (WAT) - A divided attention field sobriety test.

6
 SESSION V

PHASE ONE: VEHICLE IN MOTION

HS 178 R2/06
SESSION V

PHASE ONE: VEHICLE IN MOTION

Upon successfully completing this session, the participant will be able to:

o Identify typical cues of Detection Phase One.

o Describe the observed cues clearly and convincingly.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Overview: Tasks and Decision o Instructor-Led Presentations

B. Initial Observations: Visual Cues

Impaired Operation o Video Presentation

C. Initial Observations: Visual Cues

Impaired Operation (Motorcycles) o Video Presentation

D. Recognition and Description of

Initial Cues o Instructor-Led Demonstrations

E. Typical Reinforcing Cues of the

Stopping Sequence o Participant's Presentations

F. Recognition and Description of

Initial and Reinforcing Cues

HS 178 R2/06
 DWI DETECTION PHASE ONE: VEHICLE IN MOTION

Your first task in Phase One: Vehicle in Motion is to observe the vehicle in
operation to note any initial cues of a possible DWI violation. At this point you
must decide whether there is sufficient cause to stop the vehicle, either to conduct
further investigation to determine if the suspect may be impaired, or for another
traffic violation. You are not committed to arresting the suspect for DWI based on
this initial observation, but rather should concentrate on gathering all relevant
evidence that may suggest impairment. Your second task during phase one is to
observe the manner in which the suspect responds to your signal to stop, and to
note any additional evidence of a DWI violation.

The first task, observing the vehicle in motion, begins when you first notice the
vehicle, driver or both. Your attention may be drawn to the vehicle by such things
as:

o a moving traffic violation;

o an equipment violation;

o an expired registration or inspection sticker;

o unusual driving actions, such as weaving within a lane or moving at

slower than normal speed; or

o "Evidence of drinking" or drugs in vehicle.

If this initial observation discloses vehicle maneuvers or human behaviors that may
be associated with impairment, you may develop an initial suspicion of DWI.

Based upon this initial observation of the vehicle in motion, you must decide
whether there is reasonable suspicion to stop the vehicle. At this point you have
three choices:

o stop the vehicle;

o continue to observe the vehicle; or

o disregard the vehicle.

HS 178 R2/06 V-1

2. INITIAL OBSERVATIONS: VISUAL CUES TO DWI

Drivers who are impaired frequently exhibit certain effects or symptoms of

impairment. These include:

o slowed reactions;
o impaired judgment as evidenced by a willingness to take risks;
o impaired vision; and
o poor coordination

The next page presents common symptoms of alcohol influence. This unit focuses
on alcohol impairment because research currently provides more information about
the effects of alcohol on driving than it does about the effects of other drugs on
driving. Remember that whether the driver is impaired, the law enforcement
detection process is the same, and the offense is still DWI.

The common effects of alcohol on the driver's mental and physical faculties lead to
predictable driving violations and vehicle operating characteristics. The National
Highway Traffic Safety Administration (NHTSA) sponsored research to identify the
most common and reliable initial indicators of DWI. This research identified 24
cues, each with an associated high probability that the driver exhibiting the cue is
impaired. These cues and their associated probabilities are described in the
following Special Section, Initial Visual DWI Detection Cues.

HS 178 R2/06 V-2

They also are discussed in Visual Detection of Driving While Intoxicated, a film
sponsored by NHTSA to assist law enforcement officers to recognize DWI detection
cues. This film is included in the training video.

COMMON SIGNS OF ALCOHOL INFLUENCE

BLOOD ALCOHOL CONCENTRATION

0.03% 0.05 0.08 0.10

* * * *
* * * *
* * * *
? ? ? ?
* * * *
* * * *
* * * *
? ? ? ?
Slowed * * *
Reactions * * *
* * *
? ? ?
Increased * *
Risk * *
Taking * *
? ?
Impaired *
Vision *
?
Poor
Coordination

INITIAL VISUAL DWI DETECTION CUES

Following are 24 cues which police officers may use to detect nighttime impaired
drivers. The cues were developed from a list of more than 100 driving cues that
have been found to predict BACs of 0.08 percent or greater. Hundreds of law
enforcement officers were involved in three field studies involving more than 12,000
enforcement stops. These cues represent the most systematically developed method
available for visually predicting whether a vehicle operated at night is being driven
by a DWI driver or a sober driver.

A pocket-sized booklet, “The Visual Detection of DWI Motorists” [DOT HS 808 677]
listing these cues is available free of charge from:
National Highway Traffic Safety Administration - Impaired Driving Division
400 Seventh Street, SW., Room 5118
Washington, DC 20590

HS 178 R2/06 V-3

 VISUAL CUE DESCRIPTIONS

1. PROBLEMS MAINTAINING PROPER LANE POSITION [p=.50-.75]

A. Weaving - Weaving occurs when the vehicle alternately moves toward one
side of the roadway and then the other, creating a zig-zag course. The
pattern of lateral movement is relatively regular as one steering correction
is closely followed by another.

B. Weaving Across Lane Lines - Extreme cases of weaving when the vehicle
wheels cross the lane lines before correction is made.

C. Straddling A Lane Line - The vehicle is moving straight ahead with the
center or lane marker between the left-hand and right-hand wheels.

D. Swerving - A swerve is an abrupt turn away from a generally straight

course. Swerving might occur directly after a period of drifting when the
driver discovers the approach of traffic in an oncoming lane or discovers
that the vehicle is going off the road; swerving might also occur as an
abrupt turn is executed to return the vehicle to the traffic lane. In the
illustration below, a swerve was executed to return to a lane after a period
of drifting toward opposing traffic.

E. Turning With Wide Radius - During a turn, the radius defined by the
distance between the turning vehicle and the center of the turn in greater
than normal. The vehicle may drive wide in a curve.

F. Drifting - Drifting is a straight-line movement of the vehicle at a slight

angle to the roadway. As the driver approaches a marker or boundary (lane
marker, center line, edge of the roadway), the direction of drift might
change. As shown in the illustration, the vehicle drifts across the lane
marker into another lane, then the driver makes a correction and the
vehicle drifts back across the lane marker. Drifting might be observed
within a single lane, across lanes, across the center line, onto the shoulder,
and from lane to lane.

G. Almost Striking Object or Vehicle - The observed vehicle almost strikes

a stationary object or another moving vehicle. Examples include: passing
abnormally close to a sign, wall, building, or other object; passing
abnormally close to another moving vehicle; and causing another vehicle to
maneuver to avoid collision.

HS 178 R2/06 V-4

2. SPEED AND BRAKING PROBLEMS [p=.45-.70]

A. Stopping Problems (too far, too short, too jerky) - Stopping too far
from a curb or at an inappropriate angle. Stopping too short or beyond
limit line at an intersection. Stopping with a jerking motion or abruptly.

B. Accelerating or Decelerating Rapidly - This cue encompasses any

acceleration or deceleration that is significantly more rapid than that
required by the traffic conditions. Rapid acceleration might be
accompanied by breaking traction; rapid deceleration might be
accompanied by an abrupt stop. Also a vehicle might alternately accelerate
and decelerate rapidly.

C. Varying Speed - Alternating between speeding up and slowing down.

D. Slow Speed (10 m.p.h. + Under Limit) - The observed vehicle is being
driving at a speed that is more than 10 MPH below the speed limit.

3. VIGILANCE PROBLEMS [p=.55-.65]

A. Driving In Opposing Lanes or Wrong Way On One-Way Street - The

vehicle is observed heading into opposing or crossing traffic under one or
more of the following circumstances: driving in the opposing lane; backing
into traffic; failing to yield the right-of-way; driving the wrong way on a
one-way street. The last circumstance is illustrated below.

B. Slow Response to Traffic Signals - The observed vehicle exhibits a

longer than normal response to a change in traffic signal. For example, the
driver remains stopped at the intersection for an abnormally long period of
time after the traffic signal has turned green.

C. Slow Or Failure To Respond To Officer’s Signals - Driver is unusually

slow to respond to an officer’s lights, siren or hand signals.

D. Stopping in Lane for No Apparent Reason - The critical element in this

cue is that there is no observable justification for the vehicle to stop in the
traffic lane; the stop is not caused by traffic conditions, traffic signals, an
emergency situation, or related circumstances. Impaired drivers might
stop in the lane when their capability to interpret information and make
decisions becomes impaired. As a consequence, stopping in lane for no
apparent reason is likely to occur at intersections or other decision points.

HS 178 R2/06 V-5

 E. Driving Without Headlights At Night - The observed vehicle is being
driven with both headlights off during a period of the day when the use of
headlights is required.

F. Failure to Signal or Signal Inconsistent with Action - A number of

possibilities exist for the driver’s signaling to be inconsistent with the
associated driving actions. This cue occurs when inconsistencies such as
the following are observed: failing to signal a turn or lane change; signaling
opposite to the turn or lane change executed; signaling constantly with no
accompanying driving action; and driving with four-way hazard flashers on.

4. JUDGMENT PROBLEMS [p=.35-.90]

A. Following Too Closely - The vehicle is observed following another vehicle

while not maintaining the legal minimum separation.

B. Improper Or Unsafe Lane Change - Driver taking risks or endangering

others. Driver is frequently or abruptly changing lanes without regard to
other motorists.

C. Illegal or Improper Turn (too fast, jerky, sharp, etc.) - The driver
executes any turn that is abnormally abrupt or illegal. Specific examples
include: turning with excessive speed; turning sharply from the wrong lane;
making a U illegally; turning from outside a designated turn lane.

D. Driving on Other Than Designated Roadway - The vehicle is observed

being driven on other than the roadway designated for traffic movement.
Examples include driving: at the edge of the roadway, on the shoulder, off
the roadway entirely, and straight through turn-only lanes or areas.

E. Stopping Inappropriately In Response To Officer - The observed

vehicle stops at an inappropriate location or under inappropriate
conditions, other than in the traffic lane. Examples include stopping: in a
prohibited zone; at a crosswalk; far short of an intersection; on a walkway;
across lanes; for a green traffic signal; for a flashing yellow traffic signal;
abruptly as if startled; or in an illegal, dangerous manner.

F. Inappropriate Or Unusual Behavior (throwing objects, arguing,

etc.) - Throwing objects from the vehicle, drinking in the vehicle, urinating
at roadside, arguing without cause, other disorderly actions.

HS 178 R2/06 V-6

 G. Appearing to be Impaired - This cue is actually one or more of a set of
indicators related to the personal behavior or appearance of the driver.
Examples of specific indicators might include:

o Eye fixation
o Tightly gripping the steering wheel
o Slouching in the seat
o Gesturing erratically or obscenely
o Face close to the windshield
o Driver’s head protruding from vehicle

POST STOP CUES p > .85

1. Difficulty with motor vehicle controls

2. Difficulty exiting the vehicle
3. Fumbling with driver’s license or registration
4. Repeating questions or comments
5. Swaying, unsteady, or balance problems
6. Leaning on the vehicle or other object
7. Slurred speech
8. Slow to respond to officer/officer must repeat
9. Provides incorrect information, changes answers
10. Odor of alcoholic beverage from the driver

Ask for Visual Detection of DWI Motorists. (DOT HS 808 677).

HS 178 R2/06 V-7

 VISUAL DETECTION OF DWI MOTORCYCLISTS

NHTSA has also developed research identifying driving impairment cues for
motorcyclists (ANACAPA Sciences, DOT HS 807 839, 1993).

Excellent Cues (50% or greater probability)

o Drifting during turn or curve

o Trouble with dismount
o Trouble with balance at a stop
o Turning problems (e.g., unsteady, sudden corrections, late braking,
improper lean angle)
o Inattentive to surroundings
o Inappropriate or unusual behavior (e.g., carrying or dropping object,
urinating at roadside, disorderly conduct, etc.)
o Weaving

Good Cues (30 to 50% probability)

o Erratic movements while going straight

o Operating without lights at night
o Recklessness
o Following too closely
o Running stop light or sign
o Evasion
o Wrong way

3. DIVIDED ATTENTION

It is important to understand the effects of alcohol are exhibited in driving so that

the significance of visual cues will be recognized. Driving is a complex task
involving a number of subtasks, many of which occur simultaneously. These
include:

o steering;
o controlling the accelerator;
o signaling;
o controlling the brake pedal
o operating the clutch;
o operating to gearshift;
o observing other traffic;
o observing signal lights, stop signs & other traffic control devices; and
o making decisions (whether to stop, turn, speed up, slow down).

HS 178 R2/06 V-8

Safe driving demands the ability to divide attention among these various tasks.
"Divided attention" simply means the ability to concentrate on two or more things
at the same time. Under the influence of alcohol and/or other drugs, a driver's
ability to divide attention is impaired. As a result, the impaired driver tends to
concentrate on only the most important or critical parts of driving and to disregard
the less important parts, often creating unexpected or dangerous situations for
other drivers. Two examples were particularly evident in the video segment Visual
Detection of Driving While Intoxicated. In one instance the driver signaled for left
turn, but actually turned right. In the other, the driver remained stopped at a
green light. In each case the driver was unable to divide attention.

o The first driver was concentrating on steering, looking for the street where
he wished to turn and slowing for the turn. The driver realized that a
signal was required and actually operated the signal lever. But the driver
didn't have enough attention left to move the lever in the right direction.
Therefore he signaled left, but turned right.

o The second driver was stopped at a traffic light, but he did not have
enough attention left to react to the specific color of the light. Therefore he
did not respond to the green light.

Some of the most significant evidence from all three phases of DWI detection can be
related directly to the effects of alcohol and/or other drugs on divided attention
ability. We will return to the concept of divided attention in Session VI. Personal
Contact and Session VII. Pre-arrest Screening.

4. RECOGNIZING AND DESCRIBING INITIAL CUES

Observing the vehicle in operation is the first task in DWI detection. Proper
performance of that task requires two distinct but related abilities:

o the ability to recognize evidence of impairment; and

o the ability to describe that evidence clearly and convincingly.

It is not enough that you observe and recognize symptoms of impaired driving. You
also must be able to describe what happened so that others will have a clear mental
picture of what took place. Improving your ability to recognize and clearly describe
observational evidence requires practice.

HS 178 R2/06 V-9

 5. THE STOPPING SEQUENCE

Your second task during Phase One of the detection process is to observe the
manner in which the driver responds to your signal to stop, and to note any
additional evidence of a DWI violation.

Cues reinforcing the suspicion of DWI may be found in the stopping sequence. After
the command to stop is given, the impaired driver may exhibit additional important
evidence of DWI. These cues may include:

o an attempt to flee;
o no response;
o slow response;
o an abrupt swerve;
o sudden stop; and
o striking the curb or another object.

Some of these cues are exhibited because the stop command places additional
demands on the driver's ability to divide attention. The signal to stop creates a new
situation with which the driver must cope. Flashing emergency lights or a siren
demand and divert the driver's attention, requiring that the driver now divide
attention between driving and responding to the stop command. Stopping itself
requires the driver simultaneously to turn the steering wheel, put on the brakes,
use a turn signal, and so on. Thus the driver's task becomes more complex when
the stop command is given. An impaired driver may not be able to handle this more
complex task and additional evidence of impairment may appear.

It is your responsibility to recognize, record and convey the additional evidence of

driving impairment that may come to light during the stopping sequence. This
task, like Task One, observing the vehicle in operation, requires:

o the ability to recognize evidence of impairment; and

o the ability to describe that evidence clearly and convincingly.

Recognizing and describing the reinforcing cues of DWI that appear during the
stopping sequence requires practice.

HS 178 R2/06 V-10

 TEST YOUR KNOWLEDGE

INSTRUCTIONS: Complete the following sentences.

1. The Phase One tasks are __

__

2. Two common symptoms of impairment are:

a. _____

b. _____

3. Alcohol impairs the ability to among tasks.

4. Three cues reinforcing the suspicion of DWI which may be observed during the
stopping sequence are:

a.

b.

c.

HS 178 R2/06 V-11

 SESSION VI

PHASE TWO: PERSONAL CONTACT

HS 178 R2/06
SESSION VI

PHASE TWO: PERSONAL CONTACT

Upon successfully completing this session, the participant will be able to:

o Identify typical clues of Detection Phase Two.

o Describe the observed clues clearly and convincingly.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Overview: Tasks and Decision o Instructor-Led Presentations

B. Typical Investigation Clues of the

Driver Interview o Video Presentation

C. Recognition and Description of

Investigation Clues o Instructor-Led Demonstrations

D. Interview/Questions Techniques o Participant’s Presentations

E. Recognition and Description of Clues

Associated With the Exit Sequence

HS 178 R2/06
 PERSONAL CONTACT

OVERVIEW

DWI Detection Phase Two: Personal Contact, like Phases One and Three, com-
prises two major evidence gathering tasks and one major decision. Your first task
is to approach, observe and interview the driver while they are still in the vehicle to
note any face-to-face evidence of impairment. During this face-to-face contact you
may administer some simple pre-exit sobriety tests to gain additional information
to evaluate whether or not the driver is impaired. After this evaluation, you must
decide whether to request the driver to exit the vehicle for further field sobriety
testing. In some jurisdictions departmental policy may dictate that all drivers
stopped on suspicion of DWI be instructed to exit. It is important to note that by
instructing the driver to exit the vehicle, you still are not committed to an arrest;
this is simply another step in the DWI detection process. Once you have requested
the driver to exit the vehicle, your second task is to observe the manner in which
the driver exits to note any additional evidence of impairment.

NOTE: You may initiate Phase Two without Phase One. This may occur, for
example, at a roadblock, or when you have responded to the scene of a crash.

TASK ONE

The first task of Phase Two, observation and interview of the driver, begins as soon
as the suspect vehicle and the patrol vehicle have come to complete stops. It
continues through your approach to the suspect vehicle and involves all
conversation between you and the driver prior to the driver's exit from the vehicle.

HS 178 R2/06 VI-1

You may have developed a strong suspicion that the driver is impaired prior to the
face-to-face observation and interview. You may have developed this suspicion by
observing something unusual while the vehicle was in motion, or during the
stopping sequence, or you may have developed no suspicion of DWI prior to the
face-to-face contact. The vehicle operation and the stop may have been normal, you
may have seen no actions suggesting DWI. For example, you may have stopped the
vehicle for a equipment/registration violation, or where no unusual driving was
evident. In some cases, Phase One will have been absent. For example, you may
first encounter the driver and vehicle after a crash or when responding to a request
for motorist assistance.

Regardless of the evidence that may have come to light during Detection Phase
One, your initial face-to-face contact with the driver usually provides the first
definite indications that the driver is impaired.

DECISION

Based upon your face-to-face interview and observation of the driver, and upon your
previous observations of the vehicle in motion and the stopping sequence, you must
decide whether there is sufficient reason to instruct the driver to step from the
vehicle.

For some law enforcement officers, this decision is automatic since their agency
policy dictates that the driver always be told to exit the vehicle, regardless of the
cause for the stop. Other agencies, however, treat this as a discretionary decision,
to be based on what the officer sees, hears and smells during observation and
interview with the driver while the driver is seated in the vehicle.

If you decide to instruct the driver to exit, you must closely observe the driver's
actions during the exit from the vehicle and note any evidence of impairment.

TYPICAL INVESTIGATION CLUES: THE DRIVER INTERVIEW

Face-to-face observation and interview of the driver allows you to use three senses
to gather evidence of alcohol and/or other drug influence:

o the sense of sight;

o the sense of hearing; and
o the sense of smell.

HS 178 R2/06 VI-2

SIGHT

There are a number of things you might see during the interview that would be de-
scribable clues or evidence of alcohol and/or other drug influence. Among them are:

o bloodshot eyes;
o soiled clothing;
o fumbling fingers;
o alcohol containers;
o drugs or drug paraphernalia;
o bruises, bumps or scratches;
o unusual actions.

HEARING

Among the things you might hear during the interview that would be describable
clues or evidence of alcohol and/or other drug influence are these:

o slurred speech;
o admission of drinking;
o inconsistent responses;
o abusive language;
o unusual statements.

SMELL

There are things you might smell during the interview that would be describable
clues or evidence of alcohol and/or other drug influence. Typically these include:

o alcoholic beverages;
o marijuana;
o "cover up" odors like breath sprays;
o unusual odors.

REQUIRED ABILITIES

Proper face-to-face observation and interview of the driver demands two distinct but
related abilities:

o the ability to recognize the sensory evidence of alcohol and/or other

drug influence; and

o the ability to describe that evidence clearly and convincingly.

Developing these abilities requires practice.

HS 178 R2/06 VI-3

 PRE-EXIT INTERVIEW TECHNIQUES

A basic purpose of the face-to-face observation and interview of the driver is to

identify and gather evidence of alcohol and/or other drug influence. This is the
purpose of each task in each phase of DWI detection.

During the face-to-face observation and interview stage, it is not necessary to

gather sufficient evidence to arrest the driver immediately for DWI.

There are a number of techniques you can use while the driver is still behind the
wheel. Most of these techniques apply the concept of divided attention. They
require the driver to concentrate on two or more things at the same time. They
include both questioning techniques and psychophysical (mind/body) tasks.

These techniques are not as reliable as the standardized field sobriety tests but
they can still be useful for obtaining evidence of impairment. THESE
TECHNIQUES DO NOT REPLACE THE SFST.

Questioning Techniques

The questions you ask and the way in which you ask them can constitute simple
divided attention tasks. Three techniques are particularly pertinent:

o asking for two things simultaneously;

o asking interrupting or distracting questions; and,
o asking unusual questions.

An example of the first technique, asking for two things simultaneously, is

requesting that the driver produce both the driver's license and the vehicle
registration. Possible evidence of impairment may come to light as the driver
responds to this dual request. Be alert for the driver who:

o forgets to produce both documents;

o produces documents other than the ones requested;
o fails to see the license, registration or both while searching through
wallet or purse;
o fumbles or drops wallet, purse, license or registration;
o is unable to retrieve documents using fingertips.

HS 178 R2/06 VI-4

The second technique, asking interrupting or distracting questions, forces the driver
to divide attention between searching for the license or registration and answering
a new question. While the driver is responding to the request for license,
registration or both, you ask an unrelated question like, "Without looking at your
watch, what time is it right now?" Possible evidence of impairment may be
disclosed by the interrupting or distracting question. Be alert for the driver who:

o ignores the question and concentrates only on the license or

registration search;
o forgets to resume the search after answering the question;
o supplies a grossly incorrect answer to the question.

The third technique, asking unusual questions, is employed after you have obtained
the driver's license and registration. Using this technique, you seek verifying
information through unusual questions. For example, while holding the driver's
license, you might ask the driver, "What is your middle name?"

There are many such questions which the driver normally would be able to answer
easily, but which might prove difficult if the driver is impaired, simply because they
are unusual questions. Unusual questions require the driver to process
information; this can be especially difficult when the driver does not expect to have
to process information. For example, a driver may respond to the question about
the middle name by giving a first name. In this case the driver ignored the unusual
question and responded instead to a usual -- but unasked -- question.

ADDITIONAL TECHNIQUES

Know if there are any judicial restraints in reference to these tests.

ALPHABET

This technique requires the subject to recite a part of the alphabet. You instruct
the subject to recite the alphabet beginning with a letter other than A and stopping
at a letter other than Z. For example, you might say to a driver, "Recite the
alphabet, beginning with the letter E as in Edward and stopping with the letter P
as in Paul." This divides the driver's attention because the driver must concentrate
to begin at an unusual starting point and recall where to stop.

COUNT DOWN

This technique requires the subject to count out loud 15 or more numbers in reverse
sequence. For example, you might request a driver to, "Count out loud backwards,
starting with the number 68 and ending with the number 53." This, too, divides
attention because the driver must continuously concentrate to count backwards
while trying to recall where to stop.

HS 178 R2/06 VI-5

NOTE: This technique should never be given using starting and stopping points
that end in 0 or 5 because these numbers are too easy to recall. For example, do not
request that the driver count backwards from 65 to 50. Instead, ask the driver to
count backwards from 64 to 49.

FINGER COUNT

In this technique, the subject is asked to touch the tip of the thumb in
turn to the tip of each finger on the same hand while simultaneously
counting up one, two, three, four; then to reverse direction on the
fingers while simultaneously counting down four, three, two, one.

In each instance, note whether and how well the subject is able to
perform the divided attention task.

THE EXIT SEQUENCE

Your decision to instruct the driver to step from the vehicle usually is made after
you have developed a suspicion that the driver is impaired.* Even though that
suspicion may be very strong, usually the suspect is not yet under arrest when you
give the instruction.

How the driver steps and walks from the vehicle and actions or behavior during the
exit sequence may provide important evidence of impairment. Be alert to the driver
who:

o shows angry or unusual reactions;

o cannot follow instructions;
o cannot open the door;
o leaves the vehicle in gear;
o "climbs" out of vehicle;
o leans against vehicle;
o keeps hands on vehicle for balance.

Proper face-to-face observation and interview of a driver requires the ability to

recognize the sensory evidence of alcohol and/or other drug influence and the ability
to describe that evidence clearly and convincingly. Developing these abilities takes
practice.

*Except, however, that you may instruct a suspect to exit the vehicle as a means of
ensuring your own safety. Safety considerations take precedence over all other
considerations.
HS 178 R2/06 VI-6
 TEST YOUR KNOWLEDGE

INSTRUCTIONS: Complete the following sentences.

1. The two major evidence gathering tasks of Phase Two are

2. The major decision of Phase Two is

3. Among the describable clues an officer might see during the Phase Two
interview are these three:

a.

b.

c.

4. Among the describable clues an officer might hear during the Phase II
interview are these three:

a.

b.

c.

5. Among the describable clues an officer might smell during the Phase II
interview are these two:

a.

b.

HS 178 R2/06 VI-7

6. Three techniques an officer might use in asking questions constitute simple
divided attention tasks. These techniques are:

a.

b.

c.

7. The Count Down Technique requires the subject to

8. Leaning against the vehicle is a clue to DWI which may be observed during

HS 178 R2/06 VI-8

 SESSION VIII

CONCEPTS AND PRINCIPLES OF THE

STANDARDIZED FIELD SOBRIETY TESTS

HS 178 R2/06
SESSION VIII

CONCEPTS AND PRINCIPLES OF THE STANDARDIZED FIELD SOBRIETY

TESTS

Upon successfully completing this session, the participant will be able to:

o Discuss the development and validity of the research and the standardized
elements, clues and interpretation of the three standardized field sobriety tests.

o Discuss the different types of nystagmus and their effects on the Horizontal
Gaze Nystagmus test.

o Discuss and properly administer the three Standardized Field Sobriety Tests.

o Discuss and recognize the clues of the three Standardized Field Sobriety Tests.

o Describe in a clear and convincing fashion and properly record the results of the
three Standardized Field Sobriety Tests on a standard note taking guide.

o Discuss the limiting factors of the three Standardized Field Sobriety Tests.

CONTENTS SEGMENTS LEARNING ACTIVITIES

A. Overview: Development and Validation o Instructor-Led Presentation

B. SFST Field Validation Studies o Instructor-Led Demonstration

C. Horizontal Gaze Nystagmus o Participant Practice Session &

Demonstration
D. Vertical Gaze Nystagmus

E. Walk-and-Turn

F. Combining the Clues of the Horizontal

Gaze Nystagmus and Walk-and-Turn

G. One-Leg Stand

H. Limitations of the Three Tests

I. Taking Field Notes on the Standardized

Field Sobriety Tests

HS 178 R2/06
 OVERVIEW OF SFST
RESEARCH AND DEVELOPMENT

1. For many years law enforcement officers have utilized field sobriety tests to
determine the impairment of a person’s driving due to alcohol influence. The
performance of the person on those field sobriety tests was used by the officer to
develop probable cause for arrest and as evidence in court. A wide variety of
field sobriety tests existed and there was a need to develop a battery of
standardized valid tests.

2. Beginning in late 1975, extensive scientific research studies were sponsored by

NHTSA through a contract with the Southern California Research Institute
(SCRI) to determine roadside field sobriety tests were the most accurate. SCRI
published the following three reports:

o California: 1977 (Lab)

o California: 1981 (Lab and Field)
o Maryland, D.C., V.A., N.C., 1983 (Field)

3. SCRI traveled to law enforcement agencies throughout the United States to

select the most commonly used field sobriety tests. Six tests were used in the
initial stages of this study.

4. Laboratory research indicated that three of these tests, when administered in a

standardized manner, were a highly accurate and reliable battery of tests for
distinguishing BACs above 0.10:

o Horizontal Gaze Nystagmus (HGN)

o Walk-and-Turn (WAT)
o One-Leg Stand (OLS)

5. NHTSA analyzed the laboratory test data and found:

o HGN, by itself, was 77% accurate

o WAT, by itself, was 68% accurate
o OLS, by itself, was 65% accurate
o By combining HGN and WAT an 80% accuracy can be achieved.

6. The final phase of this study was conducted as a field validation.

o Standardized, practical and effective procedures were developed

o The tests were determined to discriminate in the field, as well as in the
laboratory.

HS 178 R2/06 VIII-1

7. The three standardized test were found to be highly reliable in identifying
subjects whose BACs were above 0.10. The results of the study unmistakably
validated the SFSTs.

SFST VALIDATION STUDIES

1. Three SFST validation studies were undertaken between 1995 and 1998:

o Colorado - 1995
o Florida - 1997
o San Diego - 1998

2. The Colorado SFST validation study was the first full field study that utilized
law enforcement personnel experienced in the use of SFSTs.

o The initial study utilized only a few experienced officers in DWI

enforcement in both a laboratory setting and field setting.

o Correct arrests decisions were made 93% of the time based on the 3-test
battery (HGN, WAT, OLS). Substantially higher than the initial study
results.

3. The Florida SFST field validation study was undertaken in order to answer
the question of whether SFSTs are valid and reliable indices of the presence
of alcohol when used under present day traffic and law enforcement conditions.

o Correct decisions to arrest were made 95% of the time based on the 3-test
battery (HGN, WAT, OLS).

o This is the third SFST field validation study that has been undertaken.
Each has shown that the SFST 3-test battery is the only scientifically
validated and reliable method for discriminating between impaired and
unimpaired drivers.

4. The San Diego SFST validation field study was undertaken because of the
nationwide trend towards lower the BAC limits to 0.08. The question to be
answered was “does SFST discriminate at BAC’s below 0.10".

o Correct arrest decisions were made 91% of the time based on the 3-test
battery (HGN, WAT, OLS) at the 0.08 level and above.

HS 178 R2/06 VIII-2

 o The results of this study provide a clear evidence of the validity of the 3-test
battery. To support arrest decisions at above or below 0.08, it strongly
suggests that the SFSTs also accurately discriminate BACs at 0.04 and
above.

OVERVIEW OF NYSTAGMUS

Nystagmus

Nystagmus is defined as an involuntary jerking of the eyes. Alcohol and certain

other drugs cause Horizontal Gaze Nytagmus.

Categories of Nystagmus

There are three general categories of nystagmus:

1. Vestibular Nystagmus is caused by movement or action to the vestibular

system.

A. Types of vestibular nystagmus:

o Rotational Nystagmus occurs when the person is spun around or rotated

rapidly, causing the fluid in the inner ear to be disturbed. If it were
possible to observe the eyes of a rotating person, they would be seen to
jerk noticeably.

o Post Rotational Nystagmus is closely related to rotational nystagmus:

when the person stops spinning, the fluid in the inner ear remains
disturbed for a period of time, and the eyes continue to jerk.

o Caloric Nystagmus occurs when fluid motion in the canals of the

vestibular system is stimulated by temperature as by putting warm
water in one ear and cold in the other.

o Positional Alcohol Nystagmus (PAN) occurs when a foreign fluid, such

as alcohol, that alters the specific gravity of the blood is in unequal
concentrations in the blood and the vestibular system.

2. Nystagmus can also result directly from neural activity:

o Optokinetic Nystagmus occurs when the eyes fixate on an object that

suddenly moves out of sight, or when the eyes watch sharply contrasting
moving images.

HS 178 R2/06 VIII-3

 Examples of optokinetic nystagmus include watching strobe lights, rotating
lights, or rapidly moving traffic in close proximity. The Horizontal Gaze
Nystagmus test will not be influenced by optokinetic nystagmus when
administered properly.

o Physiological Nystagmus is a natural nystagmus that keeps the sensory

cells of the eye from tiring. It is the most common type of nystagmus. It
happens to all of us, all the time. This type of nystagmus produces
extremely minor tremors or jerks of the eyes. These tremors are generally
too small to be seen with the naked eye. Physiological nystagmus will have
no impact on our Standardized Field Sobriety Tests, because its tremors are
generally invisible.

o Gaze Nystagmus occurs as the eyes move from the center position. Gaze
nystagmus is separated into three types:

(1) Horizontal Gaze Nystagmus occurs as the eyes move to the side. It is
the observation of the eyes for Horizontal Gaze Nystagmus that
provides the first and most accurate test in the Standardized Field
Sobriety Test battery. Although this type of nystagmus is most accurate
for determining alcohol impairment, its presence may also indicate use
of certain other drugs.

(2) Vertical Gaze Nystagmus is an involuntary jerking of the eyes (up and
down) which occurs when the eyes gaze upward at maximum elevation.
The presence of this type of nystagmus is associated with high doses of
alcohol for that individual and certain other drugs. The drugs that cause
Vertical Gaze Nystagmus are the same ones that cause Horizontal Gaze
Nystagmus.

Note: There is no drug that will cause Vertical Gaze Nystagmus that
does not cause Horizontal Gaze Nystagmus. If Vertical Gaze
Nystagmus is present and Horizontal Gaze Nystagmus is not, it could
be a medical condition.

(3) Resting Nystagmus is referred to as a jerking of the eyes as they look

straight ahead. Its presence usually indicates a pathology or high doses
of a Dissociative Anesthetic drug such as PCP. If detected, take
precautions. (OFFICER SAFETY.)

3. Nystagmus may also be caused by certain pathological disorders. They include

brain tumors and other brain damage or some diseases of the inner ear. These
pathological disorders occur in very few people and in even fewer drivers.

HS 178 R2/06 VIII-4

Medical Impairment

The examinations that you can conduct to assess possible medical impairment
include:

o Pupil size
o Resting Nystagmus
o Tracking ability

PROCEDURES

Procedures to Assess Possible Medical Impairment

Prior to administration of HGN, the eyes are checked for equal pupil size, resting
nystagmus, and equal tracking (can they follow an object together). If the eyes do
not track together, or if the pupils are noticeably unequal in size, the chance of
medical disorders or injuries causing the nystagmus is present.

Procedures of Horizontal Gaze Nystagmus Testing: The Three Clues

The test you will use at roadside is "Horizontal Gaze Nystagmus" -- an involuntary
jerking of the eyes occurring as the eyes gaze toward the side. Some jerking will be
seen if the eyes are moved far enough to the side.

1. The Lack of Smooth Pursuit (Clue Number One) - The eyes can be observed to
jerk or "bounce" as they follow a smoothly moving stimulus, such as a pencil or
penlight. The eyes of an unimpaired person will follow smoothly, i.e., a marble
rolling across a smooth pane of glass, or windshield wipers moving across a wet
windshield.

2. Distinct and Sustained Nystagmus At Maximum Deviation (Clue Number Two)

- Distinct and sustained nystagmus will be evident when the eye is held at
maximum deviation for a minimum of four seconds. People exhibit slight
jerking of the eye at maximum deviation, even when unimpaired, but this will
not be evident or sustained for more than a few seconds. When impaired by
alcohol, the jerking will be larger, more pronounced, sustained for more than
four seconds, and easily observable.

3. Onset of Nystagmus Prior To 45 Degrees (Clue Number Three) - The point at

which the eye is first seen jerking. If the jerking begins prior to 45 degrees it is
evident that the person has a BAC above 0.08, as shown by recent research.

The higher the degree of impairment, the sooner the nystagmus will be
observable.

HS 178 R2/06 VIII-5

Estimating a 45-Degree Angle

It is important to know how to estimate a 45-degree angle. How far you position the
stimulus from the suspect’s nose is a critical factor in estimating a 45-degree angle.
(i.e., If the stimulus is held 12" in front of the suspect’s nose, it should be moved 12"
to the side to reach 45 degrees. Likewise, if the stimulus is held 15" in front of the
suspect’s nose, it should be moved 15" to the side to reach 45 degrees.)

For practice, a 45-degree template can be prepared by

making a 15"-square cardboard and connecting its
opposite corners with a diagonal line.

To use this device, hold it up so that the person's nose is

above the diagonal line. Be certain that one edge of the
template is centered on the nose and perpendicular to
(or, at right angles to) the face. Have the person you are
examining follow a penlight or some other object until
suspect is looking down the 45-degree diagonal. Note
the position of the eye. With practice, you should be able
to recognize this angle without using the template.

Specific Procedures

If the suspect is wearing eyeglasses, have them removed.

Give the suspect the following instructions from a safe position. (FOR OFFICER
SAFETY KEEP YOUR WEAPON AWAY FROM THE SUSPECT):

o "I am going to check your eyes."

o "Keep your head still and follow this stimulus with your eyes only."
o "Keep following the stimulus with your eyes until I tell you to stop."

Position the stimulus approximately 12-15 inches from the suspect's nose and
slightly above eye level. Check to see that both pupils are equal in size. If they are
not, this may indicate a head injury. You may observe Resting Nystagmus at this
time, then check the suspect's eyes for the ability to track together. Move the
stimulus smoothly across the suspect's entire field of vision. Check to see if the eyes
track the stimulus together or one lags behind the other. If the eyes don't track
together it could indicate a possible medical disorder, injury, or blindness.

HS 178 R2/06 VIII-6

Check the suspect's left eye by moving the stimulus to your
right. Move the stimulus smoothly, at a speed that
requires approximately two seconds to bring the suspect's
Eye Looking
Straight Ahead

eye as far to the side as it can go. While moving the 45O

stimulus, look at the suspect's eye and determine whether

it is able to pursue smoothly. Now, move the stimulus all
the way to the left, back across suspect's face checking if
the right eye pursues smoothly. Movement of the stimulus
should take approximately two seconds out and two
seconds back for each eye. Repeat the procedure.

After you have checked both eyes for lack of smooth pursuit, check the eyes for
distinct and sustained nystagmus at maximum deviation beginning with the
suspect's left eye. Simply move the object to the suspect's left side until the eye has
gone as far to the side as possible. Usually, no white will be showing in the corner
of the eye at maximum deviation. Hold the eye at that position for a minimum of
four seconds, and observe the eye for distinct and sustained nystagmus. Move the
stimulus all the way across the suspect's face to check the right eye holding that
position for a minimum of four seconds. Repeat the procedure.

Note: Fatigue Nystagmus. This type of nystagmus may begin if a subject’s eyes are
held at maximum deviation for more than 30 seconds.

Next, check for onset of nystagmus prior to 45 degrees. Start moving the stimulus
towards the right (suspect's left eye) at a speed that would take approximately four
seconds for the stimulus to reach the edge of the suspect's shoulder. Watch the eye
carefully for any sign of jerking. When you see it, stop and verify that the jerking
continues. Now, move the stimulus to the left (suspect's right eye) at a speed that
would take approximately four seconds for the stimulus to reach the edge of the
suspect's shoulder. Watch the eye carefully for any sign of jerking. When you see
it, stop and verify that the jerking continues. Repeat the procedure. NOTE: It is
important to use the full four seconds when checking for onset of nystagmus. If you
move the stimulus too fast, you may go past the point of onset or miss it altogether.

If the suspect's eyes start jerking ADMINISTRATIVE PROCEDURES

before they reach 45 degrees, 1. CHECK FOR EYEGLASSES
check to see that some white of 2. VERBAL INSTRUCTIONS
3. POSITION STIMULUS (12-15 INCHES)
the eye is still showing on the side 4. EQUAL PUPIL SIZE AND RESTING NYSTAGMUS
closest to the ear. If no white of 5. TRACKING
the eye is showing, you either 6. LACK OF SMOOTH PURSUIT
have taken the eye too far to the 7. DIST. & SUSTAINED NYSTAGMUS @ MAX. DEV.
side (that is more than 45 8. ONSET OF NYSTAGMUS PRIOR TO 45°
9. TOTAL THE CLUES
degrees) or the person has 10. CHECK FOR VERTICAL GAZE NYSTAGMUS
unusual eyes that will not deviate
very far to the side.

HS 178 R2/06 VIII-7

NOTE: Nystagmus may be due to causes other than alcohol. These other causes
include seizure medications and some other drugs. A large disparity between the
performance of the right and left eye may indicate a medical condition.

Test Interpretation

You should look for three clues of nystagmus in each eye.

1. The eye cannot follow a moving object smoothly.

2. Nystagmus is distinct and sustained when the eye is held at maximum
deviation for a minimum of four seconds.
3. The angle of onset of nystagmus is prior to 45 degrees.

Based on the original research, if you observe four or more clues it is likely that the
suspect's BAC is above 0.10. Using this criterion you will be able to classify about
77% of your suspects accurately. This was determined during laboratory and field
testing and helps you weigh the various field sobriety tests in this battery as you
make your arrest decision.

Vertical Gaze Nystagmus

The Vertical Gaze Nystagmus test is simple to administer. During the Vertical
Gaze Nystagmus test, look for jerking as the eyes move up and are held for
approximately four seconds at maximum elevation.

1. Position the stimulus horizontally, about 12-15 inches in front of the suspect's
nose.

2. Instruct the suspect to hold the head still, and follow the object with the eyes
only.

3. Raise the object until the suspect's eyes are elevated as far as possible.

4. Hold for approximately four seconds.

5. Watch closely for evidence of jerking.

Horizontal and Vertical Gaze Nystagmus can be observed directly and does not
require special equipment. You will need a contrasting stimulus for the suspect to
follow with their eyes. This can be the tip of your index finger, penlight, or pen.
The stimulus used should be held slightly above eye level, so that the eyes are wide
open when they look directly at it. It should be held approximately 12-15 inches in
front of the nose. Remain aware of your position in relation to the suspect at all
times.

HS 178 R2/06 VIII-8

OFFICER SAFETY IS THE NUMBER ONE PRIORITY ON ANY TRAFFIC
STOP.

Procedures for Walk-and-Turn Testing

1. Instructions Stage: Initial Positioning and Verbal Instructions

For standardization in the performance of this test, have the suspect assume
the heel-to-toe stance by giving the following verbal instructions, accompanied
by demonstrations:

o "Place your left foot on the line” (real or imaginary). Demonstrate.

o "Place your right foot on the line ahead of the left foot, with heel of right foot
against toe of left foot." Demonstrate.

o "Place your arms down at your sides." Demonstrate.

o "Maintain this position until I have completed the instructions. Do not start
to walk until told to do so."

o "Do you understand the instructions so far?" (Make sure suspect indicates
understanding.)

2. Demonstrations and Instructions for the Walking Stage

Explain the test requirements, using the following verbal instructions,

accompanied by demonstrations:

o "When I tell you to start, take nine heel-to-toe steps, turn, and take nine
heel-to-toe steps back." (Demonstrate 3 heel-to-toe steps.)

o "When you turn, keep the front foot on the line, and turn by taking a series
of small steps with the other foot, like this." (Demonstrate).

o "While you are walking, keep your arms at your sides, watch your feet at all
times, and count your steps out loud."

o "Once you start walking, don't stop until you have completed the test."

o "Do you understand the instructions?" (Make sure suspect understands.)

o "Begin, and count your first step from the heel-to-toe position as 'One.'"

HS 178 R2/06 VIII-9

3. Test Interpretation

You may observe a number of different behaviors when a suspect performs this
test. Original research demonstrated that the behaviors listed below are likely
to be observed in someone with a BAC above 0.10. Look for the following clues
each time this test is given:

A. Cannot keep balance while listening to the instructions. Two tasks are
required at the beginning of this test. The suspect must balance heel-to-toe
on the line, and at the same time, listen carefully to the instructions.
Typically, the person who is impaired can do only one of these things. The
suspect may listen to the instructions, but not keep balance. Record this
clue if the suspect does not maintain the heel-to-toe position throughout the
instructions. (Feet must actually break apart.) Do not record this clue if
the suspect sways or uses the arms to balance but maintains the heel-to-toe
position.

B. Starts before the instructions are finished. The impaired person may also
keep balance, but not listen to the instructions. Since you specifically
instructed the suspect not to start walking "until I tell you to begin," record
this clue if the suspect does not wait.

C. Stops while walking. The suspect pauses for several seconds. Do not record
this clue if the suspect is merely walking slowly.

D. Does not touch heel-to-toe. The suspect leaves a space of more than one-half
inch between the heel and toe on any step.

E. Steps off the line. The suspect steps so that one foot is entirely off the line.

HS 178 R2/06 VIII-10

 F. Uses arms to balance. The suspect raises one or both arms more than 6
inches from the sides in order to maintain balance.

G. Improper turn. The suspect removes the front foot from the line while
turning. Also record this clue if the suspect has not followed directions as
demonstrated, i.e., spins or pivots around.

H. Incorrect number of steps. Record this clue if the suspect takes more or
fewer than nine steps in either direction.

Note: If suspect can't do the test, record observed clues and document the
reason for not completing the test, e.g. suspect’s safety.

If the suspect has difficulty with the test (for example, steps off the line),
continue from that point, not from the beginning. This test may lose its
sensitivity if it is repeated several times.

Observe the suspect from a safe distance and limit your movement which may
distract the suspect during the test. Always consider officer safety.

Based on original research, if the suspect exhibits two or more clues on this test
or fails to complete it, classify the suspect's BAC as above 0.10. Using this
criterion, you will be able to accurately classify 68% of your suspects.

4. Test Conditions

Walk-and-Turn test requires a designated straight line, and should be

conducted on a reasonably dry, hard, level, nonslippery surface. There should
be sufficient room for suspects to complete nine heel-to-toe steps. Note: Recent
field validation studies have indicated that varying environmental conditions
have not affected a suspect’s ability to perform this test.

The original research indicated that individuals over 65 years of age, back, leg
or inner ear problems had difficulty performing this test. Individuals wearing
heels more than 2 inches high should be given the opportunity to remove their
shoes.

5. Combined Interpretation of Horizontal Gaze Nystagmus and Walk-and-Turn

Tests

Based on the original research, combining four or more clues of HGN and two or
more clues of the Walk-and-Turn, suspects can be classified as above 0.10 BAC
80% of the time.

HS 178 R2/06 VIII-11

 Procedures for One-Leg Stand Testing

1. Instructions Stage: Initial Positioning and Verbal Instructions

Initiate the test by giving the following verbal instructions, accompanied by

demonstrations.

o "Please stand with your feet together and your arms down at the sides, like
this." (Demonstrate)

o "Do not start to perform the test until I tell you to do so."

o "Do you understand the instructions so far?" (Make sure suspect indicates
understanding.)

2. Demonstrations and Instructions for the Balance and Counting Stage

Explain the test requirements, using the following verbal instructions,

accompanied by demonstrations:

o "When I tell you to start, raise one leg, either leg, with the foot
approximately six inches off the ground, keeping your raised foot parallel to
the ground.” (Demonstrate one leg stance.)

o "You must keep both legs straight, arms at your side.”

o "While holding that position, count out loud in the following manner: “one
thousand and one, one thousand and two, one thousand and three, until told
to stop.” (Demonstrate a count, as follows: "one thousand and one, one
thousand and two, one thousand and three, etc." Officer should not look at
his foot when conducting the demonstration - OFFICER SAFETY.)

o "Keep your arms at your sides at all times and keep watching the raised
foot."

o "Do you understand?" (Make sure suspect indicates understanding.)

o "Go ahead and perform the test." (Officer should always time the 30
seconds. Test should be discontinued after 30 seconds.)

Observe the suspect from a safe distance. If the suspect puts the foot down, give
instructions to pick the foot up again and continue counting from the point at
which the foot touched the ground. If the suspect counts very slowly, terminate
the test after 30 seconds.

HS 178 R2/06 VIII-12

3. Test Interpretation

You may observe a number of different behaviors when a suspect performs this
test. The original research found the behaviors listed below are the most likely
to be observed in someone with a BAC above 0.10. Look for the following clues
each time the One-Leg Stand test is administered.

A. The suspect sways while balancing. This refers to side-to-side or back-and-

forth motion while the suspect maintains the one-leg stand position.

B. Uses arms for balance. Suspect moves arms 6 or more inches from the side
of the body in order to keep balance.

C. Hopping. Suspect is able to keep one foot off the ground, but resorts to
hopping in order to maintain balance.

D. Puts foot down. The suspect is not able to maintain the one-leg stand
position, putting the foot down one or more times during the 30-second
count.

Note: If suspect can't do the test, record observed clues and document the
reason for not completing the test, e.g. suspect’s safety.

Remember that time is critical in this test. The original research has shown a
person with a BAC above 0.10 can maintain balance for up to 25 seconds, but
seldom as long as 30.

Based on original research, if an individual shows two or more clues or fails to

complete the One-Leg Stand, there is a good chance the BAC is above 0.10.
Using that criterion, you will accurately classify 65% of the people you test as to
whether their BAC's are above 0.10.

Observe the suspect from a safe distance and remain as motionless as possible
during the test so as not to interfere. If the suspect puts the foot down, give
instructions to pick the foot up again and continue counting from the point at
which the foot touched the ground. If the suspect counts very slowly, terminate
the test after 30 seconds.

4. Test Conditions

One-Leg Stand requires a reasonably dry, hard, level, and non-slippery surface.
Suspect's safety should be considered at all times.

HS 178 R2/06 VIII-13

 The original research indicated that certain individuals over 65 years of age,
back, leg or inner ear problems, or people who are overweight by 50 or more
pounds had difficulty performing this test. Individuals wearing heels more than
2 inches high should be given the opportunity to remove their shoes.

5. Taking Field Notes on Suspects' Performance of Field Sobriety Tests

For purposes of the arrest report and courtroom testimony, it is not enough to
record the total number of clues on the three tests. The number of clues is
important to the police officer in the field because it helps determine whether
there is probable cause to arrest. But to secure a conviction, more descriptive
evidence is needed.

The officer must be able to describe how the suspect performed on the tests, and
exactly what the suspect did.

The standard note taking guide provided in this Manual is designed to help you
develop a clear description of the suspect's performance on the tests.

6. Taking Field Notes on The Eye Procedures

Equal Pupils 9 Yes 9 No
First, have subject remove glasses. Equal Tracking 9 Yes 9 No
Vertical Nystagmus 9 Yes 9 No
The section for Medical Assessment Other (i.e., Resting Nystagmus) __________________
appears at the bottom of the guide’s
front page.

o Check “Yes” or “No” box for equal pupil size.

o Check “Yes” or “No” box for equal tracking.

In the section labeled “other”, record HORIZONTAL GAZE NYSTAGMUS

any facts, circumstances, conditions,
or observations that may be relevant LEFT RIGHT

to this procedures (i.e., Resting k LACK OF SMOOTH PURSUIT

Nystagmus).
k DISTINCT AND SUSTAINED
The section on the Horizontal Gaze NYSTAGMUS AT MAXIMUM
DEVIATION
Nystagmus test appears on the
bottom of the guide's front side. k ONSET OF NYSTAGMUS
PRIOR TO 45 DEGREES

Complete the entire test for both

eyes, writing "yes" or "no" for each
nystagmus clue.

HS 178 R2/06 VIII-14

 o Write "yes" if the clue is present;
o Write "no" if the clue is not present.

In the section labeled "other," record any facts, circumstances, conditions or

observations that may be relevant to this test.

o Examples of additional evidence of impairment emerging during nystagmus

test:

- suspect unable to keep head still;

- suspect swaying noticeably;
- suspect utters incriminating statements.

o Examples of conditions that may interfere with suspect's performance of the

Horizontal Gaze Nystagmus test:

- wind, dust, etc. irritating suspect's eyes;

- visual or other distractions impeding the test (always face suspect away
from rotating lights, strobe lights and traffic passing in close proximity).

7. Taking Field Notes on Walk-and-Turn Testing

The section on the Walk-and-Turn test appears at the top of the guide's back
side.

WALK AND TURN

CANNOT KEEP BALANCE

STARTS TOO SOON

FIRST NINE STEPS SECOND NINE STEPS

STOPS WALKING

MISSES HEEL -TO- TOE

STEPS OFF LINE

RAISES ARMS

ACTUAL STEPS TAKEN

IMPROPER TURN (Describe)

CANNOT DO TEST (EXPLAIN)

OTHER:

HS 178 R2/06 VIII-15

 The first two clues, "cannot keep balance" and "starts too soon" apply only
during the instructions stage of the test. Record the number of times each of
those clues appear.

For example, if the suspect's feet "break apart" from the heel-to-toe stance twice
during the instructions stage, write "2" in the box alongside the "cannot keep
balance" clue. Similarly, if the suspect never "starts too soon," write "0" in that
box. Note: Actual steps taken is for scoring purposes only. Wrong number of
steps is the validated clue.

Don't leave boxes blank. If a particular clue never shows up, write "0" in the
corresponding box.

Record the next five clues separately for the walk down the line, and then up
the line.

A. If a suspect stops walking, record it by drawing a vertical line across the toe
of the step at which the stop occurred. Do this for the first as well as the
second nine steps. Place the letter “S” at bottom of the vertical line to
indicate stops walking.

WALK AND TURN

CANNOT KEEP BALANCE

STARTS TOO SOON

FIRST NINE STEPS SECOND NINE STEPS

STOPS WALKING

MISSES HEEL -TO- TOE

STEPS OFF LINE

RAISES ARMS

ACTUAL STEPS TAKEN

IMPROPER TURN (Describe)

CANNOT DO TEST (EXPLAIN)

OTHER:

HS 178 R2/06 VIII-16

 B. If suspect fails to touch heel-to-toe, record how many times this happens.
Draw a vertical line across the toe of the step at which the miss occurred.
Place the letter “M” at the top of the vertical line to indicate missed heel to
toe.

C. If suspect steps off the line while walking, record it by drawing a line from
the appropriate foot print at an angle in the direction in which the foot
stepped. Do it for each nine steps.

D. If suspect uses arms to balance, give some indication of how often or how
long this happens.

o Example: suspect raised arms from sides three times; place a check for
each occurrence in appropriate box.

o Example: suspect held arms away from sides during 3 through 7; place
a check for each occurrence in appropriate box.

o Example: suspect "flapped" arms continuously; make a note.

E. Record the actual number of steps taken by suspect in each direction.

For the next point, "improper turn," record a description of the turn.

If you note that the suspect "cannot perform test," indicate explicitly why
you did so.

o Example: "off line three times;"

o Example: "staggered six steps to right, nearly fell;"
o Example: "fear of injury."

At end of the test, examine each factor and determine how many clues have
been recorded. Remember, each clue may appear several times, but still
only constitutes one clue.

In the section labeled "other," record any facts, circumstances, conditions or

observations that may be relevant to this test.

o Examples of additional evidence of impairment during Walk-and-Turn

test:

- suspect verbally miscounts steps;

- suspect utters incriminating statements.

HS 178 R2/06 VIII-17

 o Examples of conditions that may interfere with suspect's performance of
the Walk-and-Turn test:

- wind/weather conditions;
- suspect's age, weight;
- suspect's footwear.

8. Taking Field Notes on the Combined Interpretation of Nystagmus and Walk-

and-Turn

By combining four or more clues of HGN with two or more clues of the WAT
test, suspects can be correctly classified as above 0.10 BAC 80% of the time.

9. Taking Field Notes on One-Leg Stand Testing ONE LEG STAND

The section on the One-Leg Stand test appears

midway down the page.

By recording when things happen as well as what

happens, you will be able to prepare a more L R
descriptive arrest report.
Sways while balancing.

You will place check marks in or near the small Uses arms to balance.
boxes to indicate how many times you observed
each of the clues. You will do this separately for Hopping.

the test on the left leg (L) or on the right leg (R).
Puts foot down.
In addition, if the
suspect puts the foot
13 20
down during the test,
you will record when Type of Footwear
it happened (write the
count on new note guide). For example, when
standing on the left leg the suspect lowered the right
foot at a count of "one thousand and thirteen", and
again at "one thousand and twenty". Your diagram
should look like the sketch to the left. You must also
pay attention to the suspect's general appearance
Type of Footwear and behavior while the test is being performed.

At end of the test, examine each factor and

determine how many distinct clues have appeared.

HS 178 R2/06 VIII-18

 IT IS NECESSARY TO EMPHASIZE THIS VALIDATION APPLIES
ONLY WHEN:

o THE TESTS ARE ADMINISTERED IN THE PRESCRIBED,

STANDARDIZED MANNER

o THE STANDARDIZED CLUES ARE USED TO ASSESS THE

SUSPECT'S PERFORMANCE

o THE STANDARDIZED CRITERIA ARE EMPLOYED TO

INTERPRET THAT PERFORMANCE.

IF ANY ONE OF THE STANDARDIZED FIELD SOBRIETY TEST

ELEMENTS IS CHANGED, THE VALIDITY IS COMPROMISED.

At end of the test, examine each factor and determine how many clues have been
recorded. Remember, each clue may appear several times, but still only constitutes
one clue.

HS 178 R2/06 VIII-19

 TEST YOUR KNOWLEDGE

INSTRUCTIONS: Complete the following sentences.

1. Walk-and-Turn is an example of field sobriety test.

2. The Walk-and-Turn requires a real or imaginary line and

3. During the stage of the Walk-and-Turn, the suspect is required to

count out loud.

4. Per the original research, the Walk-and-Turn can determine whether a

suspect's BAC is above or below 0.10, percent of the time.

5. In the Walk-and-Turn test, a suspect who steps off the line during the first 9
steps and once again during the second 9 steps and who raises arms for balance
twice during the second nine steps has produced distinct clue(s).

6. The Walk-and-Turn may not be valid when administered to persons who are
over years of age.

7. During the stage of the One-Leg Stand the suspect must maintain
balance for 30 seconds.

8. The One-Leg Stand requires that the suspect keep the foot elevated for seconds.

9. Per original research, the One-Leg Stand can determine whether a suspect's
BAC is above or below 0.10, percent of the time.

10. In the One-Leg Stand test, a suspect who sways has exhibited clue(s).

11. In the One-Leg Stand test, a suspect who raises arms, hops, and puts foot down
has exhibited clue(s).

12. The maximum number of clues for Horizontal Gaze Nystagmus that can appear
in one eye is .

13. Per original research, the HGN test can determine whether a suspect's BAC is
above 0.10, percent of the time.

14. The third clue of HGN is an onset of nystagmus prior to degrees.

HS 178 R2/06 VIII-20

 SESSION VII

PHASE THREE: PRE-ARREST SCREENING

HS 178 R2/06
SESSION VII

PHASE THREE: PRE-ARREST SCREENING

Upon successfully completing this session, the participants will be able to:

o Describe the role of psychophysical and preliminary breath tests.

o Define and describe the concepts of divided attention and nystagmus.

o Discuss the advantages and limitations of preliminary breath testing.

o Discuss the arrest decision process.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Overview: Tasks and Decision o Instructor-Led Presentations

B. Gaze Nystagmus - Definition o Instructor-Led Demonstrations

C. Horizontal Gaze Nystagmus - Definition, o Video Presentation

Concepts, Demonstrations

D. Vertical Gaze Nystagmus - Definition,

Concepts, Demonstrations

E. Divided Attention Tests: Concepts,

Examples, Demonstrations

F. Advantages and Limitations of Preliminary

Breath Testing

G. The Arrest Decision

HS 178 R2/06
 PRE-ARREST SCREENING

PHASE THREE TASKS AND DECISION

Like Phases One and Two, DWI Detection Phase Three, Pre-arrest Screening has
two major evidence gathering tasks and one major decision.

PHASE THREE:
Prearrest Screening

IS THERE PROBABLE
? CAUSE TO ARREST
THE SUSPECT FOR DWI?

Your first task in Phase Three is to administer three scientifically validated

psychophysical (field) sobriety tests. Based on these tests and on all other evidence
from Phase One and Two, you must decide whether there is sufficient probable
cause to arrest the driver for DWI. Your second task may then be to administer (or
arrange for) a preliminary breath test (PBT) to confirm the chemical basis of the
driver's impairment, if your agency uses PBTs. The entire detection process
culminates in the arrest/no arrest decision.

PSYCHOPHYSICAL TESTS

Psychophysical tests are methods of assessing a suspect's mental and physical

impairment. These tests focus on the abilities needed for safe driving: balance,
coordination, information processing and so on.

Psychophysical testing actually begins as soon as you come into face-to-face contact
with the suspect and begin the interview. Psychophysical testing continues as the
suspect steps from the vehicle and you observe the manner of the exit and walk
from the vehicle. The most significant psychophysical tests are the three
scientifically validated structured tests that you administer at roadside.

HS 178 R2/06 VII-1

PRELIMINARY BREATH-TEST

The preliminary breath test (PBT) can help to corroborate all other evidence and to
confirm your judgment as to whether the suspect is impaired. Usually PBT results
cannot be introduced as evidence against the driver in court. However, state laws
vary in this regard.

THE ARREST DECISION

The DWI detection process concludes with the arrest decision. This decision is
based on all of the evidence you have obtained during all three detection phases: on
observation of the vehicle in motion and during the stopping sequence; on face to
face observation and interview of the driver.

NYSTAGMUS

"Nystagmus" means an involuntary jerking of the eyes.

HORIZONTAL GAZE NYSTAGMUS

Horizontal Gaze Nystagmus (HGN) refers to an involuntary jerking occurring as

the eyes gaze toward the side. In addition to being involuntary the person
experiencing the nystagmus is unaware that the jerking is happening.

Involuntary jerking of the eyes becomes readily noticeable when a person is

impaired. As a person's blood alcohol concentration increases, the eyes will begin to
jerk sooner as they move to the side.

Horizontal Gaze Nystagmus is the most reliable field sobriety test. Especially when
used in combination with the divided attention tests, it will help police officers
correctly identify suspects who are impaired.

In administering the HGN test, the officer has the suspect follow the motion of a
small stimulus with the eyes only. The stimulus may be the tip of a pen or penlight,
an eraser on a pencil or your finger tip, whichever contrasts with the background.

HS 178 R2/06 VII-2

When the HGN test is administered always begin with subject's left eye. Each eye
is examined for three specific clues.

o as the eye moves from side to side, does it move smoothly or does it jerk
noticeably? (As people become impaired by alcohol, their eyes exhibit a lack
of smooth pursuit as they move from side to side.)

o when the eye moves as far to the side as possible and is kept at that position
for several seconds, does it jerk distinctly? (Distinct and sustained
nystagmus at maximum deviation is another clue of impairment.)

o as the eye moves toward the side, does it start to jerk prior to a 45-degree
angle? (Onset of nystagmus prior to 45-degrees is another clue of
impairment.)

As a person's blood alcohol concentration increases it is more likely these clues will
appear.

The maximum number of clues that may appear in one eye is three. The maximum
total number for any suspect is six. The original research shows that if four or more
clues are evident, it is likely that the suspect's blood alcohol concentration is above
0.10. With four-or-more clues present, this test is 77% accurate.

VERTICAL GAZE NYSTAGMUS

Vertical Gaze Nystagmus is an involuntary jerking of the eyes (up and down) which
occurs when the eyes gaze upward at maximum elevation. Although this type of
nystagmus was not addressed in the original research, field experience has
indicated that the presence of Vertical Gaze Nystagmus has proven to be a reliable
indicator of high doses of alcohol for that individual or certain other drugs.

HS 178 R2/06 VII-3

 DIVIDED ATTENTION TESTS

INTRODUCTION

Many of the most reliable and useful psychophysical tests employ the concept of
divided attention: they require the subject to concentrate on two things at once.
Driving is a complex divided attention task. In order to operate a vehicle safely,
drivers must simultaneously control steering, acceleration and braking; react
appropriately to a constantly changing environment; and perform many other tasks.
Alcohol and many other drugs reduce a person's ability to divide attention.
Impaired drivers often ignore the less critical tasks of driving in order to focus their
impaired attention on the more critical tasks. For example, a driver may ignore a
traffic signal and focus instead on speed control.

Even when they are impaired, many people can handle a single, focused attention
task fairly well. For example, a driver may be able to keep the vehicle well within
the proper traffic lane, as long as the road remains fairly straight. However, most
people when impaired cannot satisfactorily divide their attention to handle multiple
tasks at once.

The concept of divided attention has been applied to psychophysical testing. Field
sobriety tests that simulate the divided attention characteristics of driving have
been developed and are being used by police departments nationwide. The best of
these tests exercise the same mental and physical capabilities that a person needs
to drive safely:

o information processing;
o short-term memory;
o judgment and decision making;
o balance;
o steady, sure reactions;
o clear vision;
o small muscle control;
o coordination of limbs.

Any test that requires a person to demonstrate two or more of these capabilities
simultaneously is potentially a good psychophysical test.

Simplicity is the key to divided attention field sobriety testing. It is not enough to
select a test that just divides the subject's attention. The test also must be one that
is reasonably simple for the average person to perform when sober. Tests that are
difficult for a sober subject to perform have little or no evidentiary value.

HS 178 R2/06 VII-4

Two divided attention field sobriety tests that have proven accurate and effective in
DWI detection are the Walk-and-Turn and the One-Leg Stand. These tests are
described briefly below.

Walk-and-Turn

Walk-and-Turn is a test that has been validated through extensive research

sponsored by the National Highway Traffic Safety Administration (NHTSA). It is a
divided attention test consisting of two stages:

o Instructions Stage; and,

o Walking Stage.

In the Instructions Stage, the subject must stand with their feet in heel-to-toe
position, keep their arms at their sides, and listen to the instructions. The
Instructions Stage divides the subject's attention between a balancing task
(standing while maintaining the heel-to-toe position) and an information processing
task (listening to and remembering instructions).

In the Walking Stage the subject takes nine heel-to-toe steps, turn in a prescribed
manner, and take nine heel-to-toe steps back, while counting the steps out loud,
while watching their feet. During the turn, the subject keeps their front foot on the
line, turn in a prescribed manner, and use the other foot to take several small steps
to complete the turn. The Walking Stage divides the subject's attention among a
balancing task (walking heel-to-toe and turning); a small muscle control task
(counting out loud); and a short-term memory task (recalling the number of steps
and the turning instructions).

The Walk-and-Turn test is administered and interpreted in a standardized manner,

i.e., the same way every time. Officers administering the Walk-and-Turn test
observe the suspect's performance for eight clues:

o can't balance during instructions;

o starts too soon;
o stops while walking;
o doesn't touch heel-to-toe;
o steps off line;
o uses arms to balance;
o loses balance on turn or turns incorrectly; and,
o takes the wrong number of steps.

HS 178 R2/06 VII-5

Inability to complete the Walk-and-Turn test occurs when the suspect:

o steps off the line three or more times;

o is in danger of falling;
o cannot do the test.

Original research shows that if a suspect exhibits two or more of the clues, or
cannot complete the test, the suspect's BAC is likely to be above 0.10. This criterion
has been shown to be accurate 68 percent of the time.

ONE-LEG STAND

The One-Leg Stand test also has been validated through NHTSA's research
program. It is a divided attention test consisting of two stages:

o Instructions Stage; and,

o Balance and Counting Stage.

In the Instruction Stage, the subject must stand with feet together, keep arms at
sides, and listen to instructions. This divides the subject's attention between a
balancing task (maintaining a stance) and an information processing task (listening
to and remembering instructions.)

In the Balance and Counting Stage, the subject must raise one leg, either leg, with
the foot approximately six inches off the ground, keeping raised foot parallel to the
ground. While looking at the elevated foot, count out loud in the following manner:
"one thousand and one", "one thousand and two", “one thousand and three” until
told to stop. This divides the subject's attention between balancing (standing on one
foot) and small muscle control (counting out loud).

The timing for a thirty-second period by the officer is an important part of the
One-Leg Stand test. The original research has shown that many impaired subjects
are able to stand on one leg for up to 25 seconds, but that few can do so for 30
seconds.

One-Leg Stand is also administered and interpreted in a standardized manner.

Officers carefully observe the suspect's performance and look for four specific clues:

o sways while balancing;

o uses arms to balance;
o hops;
o puts foot down.

HS 178 R2/06 VII-6

Inability to complete the One-Leg Stand test occurs when the suspect:

o puts the foot down three or more times, during the 30-second period;
o cannot do the test.

The original research shows that, when the suspect produces two or more clues or is
unable to complete the test, it is likely that the BAC is above 0.10. This criterion
has been shown to be accurate 65 percent of the time.

PRELIMINARY BREATH TESTING

The basic purpose of preliminary breath testing (PBT) is to demonstrate the associ-
ation of alcohol with the observable evidence of the suspect's impairment. The
suspect's impairment is established through sensory evidence: what the officer
sees, hears and smells. The PBT provides the evidence that alcohol is the chemical
basis of that impairment by yielding an on-the-spot indication of the suspect's blood
alcohol concentration (BAC). The PBT provides direct indication of the BAC level.
It does not indicate the level of the suspect's impairment. Impairment varies widely
among individuals with the same BAC level.

Preliminary breath testing, like psychophysical testing, is a stage in the pre-arrest

screening of a DWI suspect. Usually the suspect is not yet under arrest when
requested to submit to the preliminary breath test. The DWI incident remains at
the investigative stage; the accusatory stage has not yet begun. The PBT result is
only one of many factors the officer considers in determining whether the suspect
should be arrested for DWI. It should never be the sole basis for a DWI arrest. The
PBT result is an important factor because it provides direct indication of alcohol
impairment. All other evidence, from initial observation of the vehicle in operation
through formal psychophysical testing, indicates alcohol impairment.

ADVANTAGES OF PBT

A PBT offers several important advantages for DWI detection. It may:

o corroborate other evidence by demonstrating that the suspicion of alcohol

impairment is consistent with the officer's observations of the suspect's
mental and physical impairment.

o confirm the officer's own judgment and help gain confidence in evaluating
alcohol impairment accurately, based on observations and psychophysical
tests. (Many officers experienced in DWI enforcement find that they rely
less and less on the PBT as their confidence in their own powers of detection
increases.)

HS 178 R2/06 VII-7

 o disclose the possibility of medical complications or impairment due to drugs
other than alcohol. (The PBT can confirm or deny that alcohol is the cause
of the observed impairment. For example, observed psychophysical
impairment coupled with a PBT result showing a very low BAC indicates an
immediate need to investigate the possibility that the suspect has ingested
a drug other than alcohol or suffers from a medical problem.)

o help to establish probable cause for a DWI arrest. (The role of the PBT in
establishing probable cause may be affected by the evidentiary value of PBT
results in your state. Consult your specific PBT law, your supervisor, or the
local prosecutor for clarification, if necessary.)

LIMITATIONS OF PBT

Preliminary breath testing may have both evidentiary limitations and accuracy
limitations. Evidentiary limitations vary with specific laws. In some states PBT
results are admissible as evidence; in other states they are not admissible. Where
the results are admissible, there may be differences in the weight or value they are
given. Consult your state PBT law, your supervisor or your local prosecutor, as
necessary, for clarification.

PBT instruments have accuracy limitations. Although all PBT instruments

currently used by law enforcement are reasonably accurate, they are subject to the
possibility of error, especially if they are not used properly. There are factors that
can affect the accuracy of preliminary breath testing devices. Some of these factors
tend to produce "high" test results; others tend to produce "low" results.

There are two common factors that tend to produce high results on a PBT.

o Residual mouth alcohol. After a person takes a drink, some of the alcohol
will remain in the mouth tissues. If the person exhales soon after drinking,
the breath sample will pick up some of this left-over mouth alcohol. In this
case, the breath sample will contain an additional amount of alcohol and the
test result will be higher than the true BAC.

It takes approximately 15 minutes for the residual alcohol to evaporate from

the mouth.

The only sure way to eliminate this factor is to make sure the suspect does
not take any alcohol for at least 15 to 20 minutes before conducting a breath
test. Remember, too, that most mouthwashes, breath sprays, cough syrups,
etc., contain alcohol and will produce residual mouth alcohol. Therefore, it
is always best not to permit the suspect to put anything in their mouth for
at least 15 to 20 minutes prior to testing.

HS 178 R2/06 VII-8

 o Breath Contaminants. Some types of preliminary breath tests might react
to certain substances other than alcohol. For example, substances such as
ether, chloroform, acetone, acetaldehyde and cigarette smoke conceivably
could produce a positive reaction on certain devices. If so, the test would be
contaminated and its result would be higher than the true BAC. Normal
characteristics of breath samples, such as halitosis, food odors, etc., do not
affect accuracy.

There are two common factors that tend to produce low PBT results.

o Cooling of the breath sample. If the captured breath sample is allowed to

cool before it is analyzed, some of the alcohol vapor in the breath may turn
to liquid and precipitate out of the sample. If that happens, the subsequent
analysis of the breath sample will produce a low BAC result.

o The composition of the breath sample. Breath composition means the mix-
ture of the tidal breath and alveolar breath. Tidal breath is breath from the
upper part of the lungs and the mouth. Alveolar breath is deep lung breath.
Breath testing should be conducted on a sample of alveolar breath, obtained
by having the subject blow into the PBT instrument until all air is expelled
from the lungs.

Radio frequency interference (RFI) can produce either high or low test results, or
can prevent a breath test device from producing any result. Care should be
exercised when utilizing a PBT around radio equipment

THE ARREST DECISION

Your arrest/no arrest decision is the culmination of the DWI detection process.
Your decision is based on all the evidence you have accumulated during each
detection phase.

PHASE ONE:
o initial observation of vehicle in motion;
o observation of the stop.

PHASE TWO:
o face-to-face observation and interview;
o observation of the exit.

PHASE THREE:
o SFSTs;
o preliminary breath tests.

Your decision involves a careful review of each of the observations you have made.

HS 178 R2/06 VII-9

Conduct a "mental summary" of the evidence collected during vehicle in motion,
personal contact and pre-arrest screening. If all of the evidence, taken together,
establishes probable cause to believe that DWI has been committed, you should
arrest the suspect for DWI. Under no circumstances should you charge the suspect
with a lesser offense instead of DWI if there is probable cause to believe that DWI
has been committed. Any reduction of DWI to a lesser charge is the responsibility
of the prosecutor or judge.

HS 178 R2/06 VII-10

TEST YOUR KNOWLEDGE

INSTRUCTIONS: Complete the following sentences.

1. The two major evidence gathering tasks of Phase Three are

2. The major decision in Phase Three is

3. The entire DWI detection process culminates in

4. Divided attention tests require the subject to

HS 178 R2/06 VII-11

5. Among the mental and physical capabilities a person needs to drive safely are
these four:

a.

b.

c.

d.

6. The two stages of the Walk-and-Turn are:

a.

b.

7. The two stages of the One-Leg Stand are:

a.

b.

8. The purpose of PBT is

9. Two factors that produce high results on a PBT are:

a.

b.

10. Two factors that produce low results on a PBT are:

a.

b.

HS 178 R2/06 VII-12

 SESSION IX

TEST BATTERY DEMONSTRATIONS

HS 178 R2/06
SESSION IX

TEST BATTERY DEMONSTRATIONS

Upon successfully completing this session, the participant will be able to:

o Demonstrate the appropriate administrative procedures for the Standardized

Field Sobriety Testing Battery.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Live Classroom Demonstrations o Instructor-Led Demonstration

B. Video Demonstration o Participant Demonstration

o Video Presentation

HS 178 R2/06
 TEST BATTERY DEMONSTRATIONS

In this session, you will have the opportunity to observe several demonstrations of

the three Standardized Field Sobriety Tests. Your instructors will conduct some of

these demonstrations. Other demonstrations will be provided on video.

HS 178 R2/06 IX-1

 SESSION X

"DRY RUN" PRACTICE SESSION

HS 178 R2/06
SESSION X

"DRY RUN" PRACTICE SESSION

Upon successfully completing this session, the participant will be able to:

o Demonstrate the proper administration of the three Standardized Field

Sobriety Tests.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Procedures and Group Assignments o Instructor-Led Presentation

B. Live Administration of SFST Battery o Instructor-Led Demonstration

C. Hands On Practice o Participant Practice Session

HS 178 R2/06
 "DRY RUN" PRACTICE SESSION

In this session, you will work with other participants, taking turns administering

the Standardized Field Sobriety Tests to each other. When you are not

administering a test or serving as the test subject, you will be expected to observe

the test administrator and subsequently help critique their performance.

The Student Performance Checklist (shown on the next two pages) should be used

to help you monitor a fellow student's performance as a test administrator.

HS 178 R2/06 X-1

 PARTICIPANT PERFORMANCE CHECKLIST

STANDARDIZED FIELD SOBRIETY TEST BATTERY

Student Name: Date:

I. HORIZONTAL GAZE NYSTAGMUS

1. Have subject remove glasses if worn.

2. Stimulus held in proper position (approximately 12"-15" from nose,

slightly above eye level).

3. Check for equal pupil size and resting nystagmus.

4. Check for equal tracking.

5. Smooth movement from center of nose to maximum deviation in

approximately 2 seconds and then back across subject's face to

maximum deviation in right eye, then back to center. Check left eye,

then right eye. (Repeat)

6. Eye held at maximum deviation for a minimum of four seconds (no

white showing). Check left eye, then right eye. (Repeat)

HS 178 R2/06 X-2

 7. Eye moved slowly (approximately 4 sec.) from center to 45 angle.

Check left eye, then right eye. (Repeat)

8. Check for Vertical Gaze Nystagmus. (Repeat)

II. WALK-AND-TURN

1. Instructions given from a safe position.

2. Tells subject to place feet on line in heel-to-toe manner (left foot behind

right foot) with arms at sides and gives demonstration.

3. Tells subject not to begin test until instructed to do so and asks if

subject understands.

4. Tells subject to take nine heel-to-toe steps and demonstrates.

5. Explains and demonstrates turning procedure.

6. Tells subject to return with nine heel-to-toe steps.

HS 178 R2/06 X-3

 7. Tells subject to count steps out loud.

8. Tells subject to look at feet while counting.

9. Tells subject not to raise arms from sides.

10. Tells subject not to stop once they begin.

11. Asks subject if all instructions are understood.

III. ONE-LEG STAND

1. Instructions given from a safe position.

2. Tells subject to stand straight, place feet together, and hold arms at

sides.

3. Tells subject not to begin test until instructed to do so and asked if

subject understands.

4. Tells subject to raise one leg, either leg, approximately 6" from the

ground, keeping raised foot parallel to the ground, and gives

demonstration.

HS 178 R2/06 X-4

 5. Tells subject to keep both legs straight and to look at elevated foot.

6. Tells subject to count by thousands in the following manner: one

thousand and one, one thousand and two, one thousand and three,

until told to stop, and gives demonstration.

7. Checks actual time subject holds leg up.

Instructor:

HS 178 R2/06 X-5

 SESSION XIV-A

"TESTING SUBJECTS" PRACTICE: SECOND SESSION

(OPTION TWO ONLY)

HS 178 R2/06
SESSION XIV-A

"TESTING SUBJECTS" PRACTICE: SECOND SESSION (OPTION TWO ONLY)

Upon successfully completing this session, the participant will be able to:

o Properly administer the SFST's.

o Properly observe and record subject's performance utilizing the standard

notetaking guide.

o Properly interpret the subject's performance.

o Proper use and maintenance of SFST Field Arrest Log.

CONTENT SEGMENTS LEARNING ACTIVITIES

A. Procedures o Instructor-Led Presentation

B. Practical Exercise (OPTION TWO ONLY) o Video Presentations

C. Session Wrap-Up o Instructor-Led Discussion

HS 178 R2/06
 "TESTING SUBJECTS" PRACTICE: SECOND SESSION

During this session, if you are attending the OPTION TWO version of this training
program, you will be administering the SFSTs to other participants and viewing
videoed volunteers who have consumed alcoholic beverages. Some of these
volunteers will have BACs above 0.08. Others will be below that level. You will
carefully note and record the volunteers' performance, and attempt to distinguish
those "0.08 and above" from those "below 0.08".

You will record your results on the SFST Field Arrest Log. Participants trained
using this option must continue to maintain this log following their training.

HS 178 R2/06 XIV-A-1

 SFST FIELD ARREST LOG

BAC Arrest Measured

Date Name HGN WAT OLS +/- .08 Not Arrest BAC Remarks

HS 178 R2/06 XIV-A-2

ATTACHMENT(S)
 ATTACHMENT B

The Detection of DWI at BACs Below 0.10

Final Report

Submitted to:
U.S. DEPARTMENT OF TRANSPORTATION
NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION

Jack Stuster, PhD, CPE

Project Director

12 September 1997

ANACAPA SCIENCES, INC.

P.O. Box 519
Santa Barbara, California 93102
(805) 966-6157
 Executive Summary

This report documents the research activities and presents the results of a
study conducted for the National Highway Traffic Safety Administration (NHTSA)
to identify driving and other behavioral cues that are associated with blood alcohol
concentrations (BACs) below the 0.10 level. The ultimate objective of the research
has been to develop training materials to assist law enforcement officers in the
accurate detection of motorists who are driving while impaired (DWI).

Description of the Research

The research and development project was composed of 13 major project
tasks, conducted in two phases. During Phase I, a work plan was developed to
guide all subsequent tasks, a comprehensive review of the low BAC literature was
performed, interviews were conducted with DWI experts from across the United
States, a data base of low BAC arrest reports was assembled, and two field studies
were conducted. The analysis of archival, interview, arrest report, and field data
collected by observers led to the identification of 34 driving cues and 10 post-stop
cues for further evaluation.

Five law enforcement agencies participated in the second of the field studies,
known as the preliminary field study, by recording the driving and post-stop cues
observed for all enforcement stops, regardless of the disposition of the stop; the
BACs of all drivers who exhibited objective signs of having consumed alcohol also
were recorded. By collecting data about all enforcement stops that were made, it
was possible to calculate the proportions of the stops in which specific cues were
found in association with various BAC levels. All archival, interview, and field
study data were analyzed, and recommendations for draft training materials were
developed, as the final Phase I task.

A draft DWI detection guide, training booklet, and training video were
developed based on the results of the preliminary field study; the materials included
24 driving and 10 post-stop cues. Law enforcement agencies representing 11 of the
15 states with 0.08 BAC limits for DWI were recruited to participate in the Phase II
validation study. Participating officers reviewed the video and printed training
materials, then completed a data collection form following every enforcement stop
made, regardless of the disposition of the stop; the same form was used as in the
preliminary field study, conducted previously. The validation study data were
analyzed and a final version of the training materials, and this technical report,
were prepared as the final Phase II project tasks.

Data were collected during more than 12,000 enforcement stops during this
research project. The stops were made by several hundred participating officers,
representing more than 50 law enforcement agencies from across the United States.

-- i --
Results
The results of the preliminary field study largely supported the 20 cues at the
0.08 BAC level that were presented on the original NHTSA DWI detection guide,
which was developed in 1980 for the 0.10 BAC level. However, no cues were found
that reliably predicted BACs below 0.08; that is, the cues that are key predicotrs of
DWI at the 0.08 BAC level failed to emerge with useful probabilities at BAC levels
below 0.08. The results of the Phase II validation study further confirmed the key
cues that were contained in the original NHTSA guide, a few additional driving
cues, and the 10 post-stop cues. The DWI driving cues were presented in functional
categories in both the printed materials and the training video: Problems
Maintaining Proper Lane Position, Speed and Braking Problems, Vigilance
Problems, and Judgment Problems.

Slight modifications were made to the training materials, based on the

results of the Phase II validation study. The final version of the DWI detection
guide is reproduced below.

DWI DETECTION GUIDE

Weaving plus any other cue: p = at least .65
Any two cues: p = at least .50

PROBLEMS MAINTAINING PROPER LANE POSITION POST STOP CUES p>.85

• Weaving • Weaving across lane lines p=.50-.75 • Difficulty with motor vehicle controls
• Straddling a lane line • Swerving • Difficulty exiting the vehicle
• Turning with a wide radius • Drifting • Fumbling with driver’s license or registration
• Almost striking a vehicle or other object • Repeating questions or comments
• Swaying, unsteady, or balance problems
SPEED AND BRAKING PROBLEMS p=.45-.70 • Leaning on the vehicle or other object
• Stopping problems (too far, too short, or too jerky) • Slurred speech
• Accelerating or decelerating for no apparent reason • Slow to respond to officer/officer must repeat
• Varying speed • Slow speed (10+ mph under limit) • Provides incorrect information, changes answers
• Odor of alcoholic beverage from the driver
VIGILANCE PROBLEMS p=.55-.65
• Driving in opposing lanes or wrong way on one-way
• Slow response to traffic signals
• Slow or failure to respond to officer’s signals
• Stopping in lane for no apparent reason
• Driving without headlights at night*
• Failure to signal or signal inconsistent with action* *p>.50 when combined with any other cue:
• Driving without headlights at night
JUDGMENT PROBLEMS p=.35-.90 • Failure to signal or signal inconsistent with action
• Following too closely
• Improper or unsafe lane change The probability of detecting DWI by random traffic
• Illegal or improper turn (too fast, jerky, sharp, etc.) enforcement stops at night has been found to be about
• Driving on other than the designated roadway three percent (.03).
• Stopping inappropriately in response to officer
• Inappropriate or unusual behavior (throwing, arguing, etc.)
• Appearing to be impaired

-- ii --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

DWI Cues At BACs Below 0.10

A Review of the Literature

The purpose of this review is to prepare information for the research team
concerning the determination and validation of visual cues for the detection of
motorists who are driving while impaired (DWI) with blood alcohol concentrations
(BACs) below 0.10.

BACKGROUND
An emphasis on DWI enforcement during the past decade has been a factor in
the significant improvement in traffic safety, as represented by declining fatal and
alcohol-involved crash rates. Despite the significant improvements in traffic safety
during the past 30 years, particularly during the past decade, more than 40,000
people still perish each year as a result of motor vehicle crashes. The current US
traffic fatality rate amount to a daily average of about 126 people – the equivalent
of a Boeing 727 crashing every day of the year.

The economic losses from alcohol involved crashes are staggering at an

estimated $21 to $24 billion annually (for property damage alone) (Miller, 1992). In
1990, the combined cost of all traffic collisions was $137.5 billion, including 28
million vehicles damaged, 5.4 million people injured, and 44,531 lives lost (Blincoe
& Faigin, 1992).

A reduction in the number of alcohol-involved crashes and the number of

alcohol-impaired drivers on the road is a top priority. Numerous studies indicate
that when DWI enforcement levels are increased, the number of alcohol involved
collisions decrease (Hause, Chavez, Hannon, Matheson, 1977; Voas & Haus, 1987;
Blomberg, 1992). However, many officers are unable to identify legally impaired
drivers from their driving behavior, or even during the brief interview customary at
a sobriety checkpoint. For example, in the Netherlands, as many as 32 percent of
drivers with BACs above .05 might escape detection at checkpoint, when officers
have the advantage of a face-to-face exchange (Gundy & Verschuur, 1986).

There are at least two clear solutions to the low BAC DWI detection problem:
1) Random Breath Testing (RBT) to objectively detect drivers operating above the
legal limit; and, 2) increased officer sensitivity to behavioral cues exhibited at lower
BAC levels. Although the RBT method is operating effectively in Australia (McCaul
& McLean, 1990), it is probably not an appropriate program for the United States.
Fourth Amendment rights currently prevent random breath testing; for example,
testing only can occur at a sobriety checkpoint after probably cause has been
established (Voas, 1991). Thus, the most likely solution to improving detection of
low BACs is to improve the DWI detection ability of law enforcement officers.

-- B - 1 --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

In 1980, Harris et al. conducted NHTSA sponsored research to determine the

behavioral cues for on-the-road detection of DWI. The final product of this Anacapa
Sciences’ study was a DWI Detection Guide providing 20 visual cues commonly
exhibited by impaired drivers with a BAC equal to or greater than 0.10. The Guide
provides the probability for each cue of discriminating between Driving While
Impaired (DWI) and Driving While Sober (DWS). The DWI Detection Guide and
supporting training materials are part of the DWI Detection and Standardized
Field Sobriety Testing course currently distributed by NHTSA (NHTSA, 1990).
Surprisingly, although there has been a limited evaluation of the DWI Detection
Guide (Vingilis et al., 1983), the only additional research of this type that has been
performed since 1980 was a NHTSA sponsored study to develop a motorcycle DWI
detection guide (Stuster, 1993).

It is legitimate to question whether a cue guide calibrated for the 0.08 level
would appear very similar if not identical to the DWI detection guide developed
nearly 20 years ago by Anacapa Sciences. A new, lower BAC limit DWI detection
guide might ultimately appear similar to the old guide, but the research is
important for at least three reasons.

1. The research that supported the development of the DWI Detection Guide
was conducted 18 years ago. Many things have changed considerably since
the late 1970s. It is not unreasonable to suspect that some fundamental
changes might be reflected in the behavioral cues associated with driver
impairment. And, there might be behaviors that correlate more closely
with lower than higher BACs.

2. At the very least, a periodic reprise of a research and development effort is

warranted if the work involved important public policy and enforcement
implications. The DWI Detection Guide and training program have not
been reviewed or revised since they were developed. Increased awareness
of DWI issues and public support for DWI enforcement in recent years
contribute to the need to upgrade and make current an important decision
aid and training program that is used by law enforcement personnel from
across the U.S.

3. It is essential for researchers to view the issue of DWI detection form the
perspective of an officer on patrol. A patrol officer wants to know the
likelihood that a specific driver behavior is indicative of DWI at the (new)
0.08 level or above, or at the 0.04 level or above. The “or above” is
important because as the BAC level is reduced the probability that a given
cue is predictive of DWI rises – because all of the or aboves are included in
the calculation. From the officer’s perspective (in an 0.08 jurisdiction) it is
usually irrelevant if the motorist is 0.08, 0.10, or some higher value – it is
only important to determine that the motorist is 0.08 or above.

-- B - 2 --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

Although the modal BAC limit for DWI continues to be 0.10 in the United
States, there is a definite trend towards lowering the limit. When the current
project started in 1993, only five states had adopted a 0.08 percent legal limit, but
by the conclusion of the research the number of states with a 0.08 limit had
increased to 15. Further, the Commercial Motor Vehicle Safety Act of 1986
established a nationwide maximum BAC of 0.04 percent for all commercial drivers.
In addition, several states have adopted a zero tolerance statute or a 0.02 percent
BAC limit for youthful drivers. Studies that suggest low officer DWI detection
rates, and improved low BAC detection when using passive alcohol sensors (Kiger et
al., 1983; Jones et al., 1985: Vingilis and Gingilis, 1985), suggest the need for a DWI
detection guide for levels below 0.10 percent BAC.

RELEVANT RESEARCH
The trend of lowering BAC limits is a reflection of the growing body of evidence
that alcohol begins to impair nervous function at BAC levels below 0.10 percent.
Moskowitz and Robinson (1988) conducted a comprehensive literature review
concerning the effects of alcohol on driving behavior, emphasizing the BACs at
which impairment begins. A majority of studies found impairment at low BACs
(below 0.07). Many studies found impairment at the 0.04 level and below.

Moskowitz and Robinson computed BACs for all studies, even those that
included BAC data in the original report. Often these calculations resulted in
higher BACs than were reported in the original study, probably because the older
devices were inaccurate. The calculations also allowed for gender differences (by
taking into account the different percentages of body water in females and males).
If anything, the calculations performed by Moskowitz and Robinson lead to an over-
estimation of BAC level. If this is the case, the impairments they report at various
BAC levels actually might occur at lower BACs than reported later in this review.

In the Moskowitz study, factors were grouped into behavioral categories

pertinent to driving. The following categories were affected at 0.05 percent BAC.

• Reaction time • Tracking

• Divided attention • Information processing
• Visual functions • Perception

Driving behaviors that showed impairment at 0.08 percent to as low as 0.03

percent included:

• Steering • Gear changing

• Braking • Speed judgment
• Speed control • Distance judgment
• Lane tracking

-- B - 3 --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

In addition, tasks requiring divided attention showed impairment at BACs as

low as 0.02 percent. These driver behaviors are listed in the table presented at the
end of this section; the table provides a comprehensive inventory of all DWI cues
identified during the current review.

Although the Moskowitz and Robinson review is the most extensive source of
information available about driver impairment at various BAC levels, several other
studies identify potential cues for DWI detection. In an Anacapa Sciences’ study
conducted for the Insurance Institute for Highway Safety, Casey and Stuster (1982)
identified the following 12 risky driving behaviors of both automobile and
motorcycle operators.

• Running stop sign or traffic light

• Unsafe passing due to oncoming traffic
• Unsafe turn in front of oncoming or opposing traffic
• Following too closely
• Unsafe lane change or unsafe merging
• Weaving through traffic
• Crossing a double line in order to pass
• Passing on the right
• Excessive speed for conditions
• Improper turn
• Splitting traffic
• Stunts

Similarly, Treat et al. (1980), in a study of risky driving actions and their
involvement in traffic collisions, identified the following 13 Unsafe Driving Actions.

• Pulling out in front of traffic

• Following behavior
• Speeding: Absolute/Over limit
• Speeding: Relative/For traffic conditions
• Turning in front of oncoming traffic
• Running stop sign or light
• Changing lanes or merging in front of traffic
• Driving left of center or on centerline
• Passing unsafely
• Driving off road to right
• Backing unsafely
• Turning too wide or too sharp
• Turning from wrong lane

-- B - 4 --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

Several of these unsafe driving actions also have been identified as indicators of
driving while impaired in the Harris et al (1980) study: following too closely, fast
speed (deleted from the final version of the DWI Detection Guide), failing to respond
to traffic signals or signs, and driving into opposing or crossing traffic.

Additionally, several studies suggest stopping method as a primary difference

between DWI and unimpaired driving (Attwood et al., 1980; Bragg et al., 1981;
Compton, 1985). Differences included braking sooner and stopping jerkily when
under the influence of alcohol.

In a study developing and validating the sobriety field test battery, Tharp,
Burns, and Moskowitz (1981) reported the reasons for stopping suspected alcohol
impaired drivers. The most common reasons were traffic infractions (e.g., speeding,
failing to stop) rather than non-infraction driving behaviors such as weaving or
drifting. There is significant overlap between the behaviors reported by Tharp et al.
(1981) and the DWI on-the-road detection cues identified by Harris et al. (1980).

In a study evaluating screening procedures for police officers at sobriety

checkpoints, cues noticed by officers were correlated with the BAC levels of the
drivers. Compton (1985) found significant differences in stopping behavior. In
general, drivers stopped smoothly at low BAC levels (0-0.04) and “jerkily” at higher
BAC levels (0.10-0.15). Drivers with a low BAC did not serve, those with higher
BACs (greater than 0.10) did. Cues identified by Compton that related to driving
and stopping behaviors, and personal appearance, are presented in the
comprehensive table at the end of this review. The cues identified in the Compton
study include personal appearance variables not previously identified in the 1980
Harris et al. study. These cues include:

• Odor of alcohol
• Face flushed
• Speech slurred
• Eyes dilated
• Demeanor
• Hair disheveled
• Poor dexterity
• Clothes disheveled

Of these personal appearance variables, odor of alcohol, face flushed, and eyes
dilated appear to be the most promising for DWI detection at low BAC levels.

-- B - 5 --
Anacapa Sciences, Inc.: Appendix B
DWI Detection at BACs Below 0.10

CONCLUSIONS
The objective of the current study is to develop an appropriate set of behaviors
that can be used by field officers to accurately identify motorists who are driving
while impaired at the 0.08 level, and to determine if cues are available that predict
0.04 and 0.02 BAC levels. No sources were identified that specifically identify
behavioral cues for alcohol impairment at the lower levels. However, a table of
potentially applicable behaviors has been prepared, based on a comprehensive
review of the literature. This list, presented in the following table, includes all
behaviors previously discussed in this review, and shows the considerable
agreement among the studies. The behaviors identified here later will be combined
with cues identified during interviews with DWI patrol experts, and from the
archival research. The resulting comprehensive inventory of DWI cues then will be
used to develop data collection forms for the first of the field studies.

-- B - 6 --
 ATTACHMENT B

A Colorado Validation Study

of the

Standardized Field Sobriety Test (SFST) Battery

Final Report Submitted to

Colorado Department of Transportation
November 1995

Marcelline Burns, Ph.D. Ellen W. Anderson, Deputy

Southern California Research Institute Pitkin County Sheriff’s Office
Los Angeles, California Aspen, Colorado

This report was funded by the Office of Transportation Safety, Colorado

Department of Transportation
(utilizing National Highway Traffic Safety Administration funds, under Project
Number 95-408-17-05).

HS 178 R2/06
 I. Introduction

A battery of standardized field sobriety Also, officer experience with the SFSTs is
tests (SFSTs), which was developed under key to the skill and confidence with which they
National Highway Traffic Safety (NHTSA) use them as a basis for their decisions. Thus
funding during the 1970's, is now used by it is important to understand that the officers
police officers nationwide. Traffic officers in who participated in the SCRI studies had not
fifty states, who have been trained in been trained with the SFSTs until just prior to
standardized administration of the tests, the experiments. They had not had
routinely use them and incorporate their opportunity and time to gain skill or to
observations of drivers’ test performance into develop confidence in the tests. In contrast,
their arrest or release decisions. Defense many of the officers who now use and testify
attorneys, however, often challenge the about the tests have been using them
admissibility of court testimony about the test regularly for ten or more years, and it is
battery. reasonable to assume they have gained skill
Roadside decisions are a critical and to expect that their decisions based on the
components of alcohol-and-driving tests may be more accurate than those of the
enforcement, and, therefore, of traffic safety. officers during the initial research.
Because the SFSTs aid officers in the often- The question to be addressed in 1995 by
difficult task of identifying alcohol-impaired agencies, officers and the courts is, “How
drivers, it is likely that the tests have accurate are the arrest decisions which are
contributed in some unknown measure to the made by experienced, skilled officers under
significant decline in alcohol-related fatalities roadside conditions when they rely on
over the last decade. Given that they have SFSTs?”. A broadly applicable answer cannot
exerted a positive impact on traffic safety, it is be found in laboratory research. It requires
important to resolve questions about their field data; i.e., information about real-world
validity and reliability, to maintain their arrest decisions by officers trained by NHTSA
credibility, and to preserve them as a roadside guidelines to administer the SFSTs.
tool. The Colorado Department of
Because court arguments about SFSTs Transportation funded a 1995 study to obtain
focus largely on the research conducted at the such data. Through a grant to the Pitkin
Southern California Research Institute (SCRI) County Sheriff’s Office and with the
and because that research is sometimes cooperative effort of seven Colorado law
misrepresented or misunderstood, it is enforcement agencies, records were collected
necessary first to clarify its purpose. Two from drivers tested with the SFSTs at
large-scale laboratory experiments were roadside. The seven agencies were:
conducted for the purpose of identifying and Aspen Police Department (APD)
standardizing a “best” set of tests (Burns and Basalt Police Department (BPD)
Moskowitz, 1977; Tharp, burns and Boulder County Sheriff’s Office (BCSO)
Moskowitz, 1981). Although it clearly is Colorado State Patrol (CSP)
relevant at this point in time to inquire Lakewood Police Department (LPD)
whether the methods of those experiments Pitkin County Sheriff’s Office (PCSO)
were scientifically sound, it should be Snowmass Village Police Dept (SVPD)
recognized that the laboratory data are now
only indirectly enlightening about current With information drawn from impaired-
roadside use of the tests. In particular, note driving records, a data base was created and
that controlled laboratory conditions are less analyzed at the Souther California Research
variable and, therefore, may be less Institute.
challenging than the highly varied conditions
which officers routinely encounter in the field.

HS 178 R2/06 1
 Technical Summary

In the State of Colorado, motor vehicle volunteered for the study provided the records
operators are subject to arrest if they are (N=305) from every administration of the
found to be driving with a blood alcohol SFSTs.
concentration (BAC) of 0.05% or higher. At Using only the standardized 3-test
BACs of 0.05% or higher but less than 0.10%, battery (Walk-and-Turn, One-Leg Stand,
they are charged with Driving While Ability Horizontal Gaze Nystagmus), officers seldom
Impaired (DWAI). At BACs of 0.10% and erred when they decided to arrest a driver.
higher, the charge is Driving Under the
Influence (DUI). These statutes reflect the Breath or blood specimens confirmed that
evidence from both epidemiological and 93% of the arrested drivers were above 0.05%
laboratory studies of alcohol impairment of BAC.
driving skills.
It is the responsibility of law enforcement
officers to detect and arrest alcohol-influenced Officers were more likely to err on the
drivers in accordance with these statutory side of releasing drivers than on the side of
limits. In an efforts to meet that objective, incorrectly arresting drivers. Given the
police officers, not only in Colorado but in all difficulty of the task which confronts officers
fifty of the United States, rely on a battery of at roadside, in particular with alcohol-tolerant
standardized field sobriety tests (SFSTs). individuals, the finding that approximately
Observations of drivers’ performance of the one-third of the released drivers should have
tests, together with driving pattern, been arrested is not unexpected. However, it
appearance and manner, odor of alcohol, and is important to note that officers’ decisions to
other signs, underlie officers’ arrest and release were correct two-thirds of the time.
release decisions.
To be genuinely useful, roadside tests Overall, 86% of the officers’ decisions to
must be valid and reliable; i.e., they must arrest or release drivers who provided blood
measure changes in performance associated or breath specimens were correct.
with alcohol and they must do it consistently.
To the extent that they meet the validity and
reliability criteria, they can be expected to
contribute to traffic safety by increasing the It is concluded that the SFSTs are valid
likelihood that alcohol-impaired drivers will be tests; i.e., they serve as indices of the presence
removed from the roadway by arrest. of alcohol at impairing levels. The study
Importantly, they also will further serve the design did not support an examination of test-
driving public’s interest by decreasing the retest reliability. It should be noted, however,
likelihood that a driver who is not alcohol- that the test battery appears to have served
impaired will be mistakenly detained or equally well across agencies and officers,
arrested. Thus, the validity and reliability of strongly suggesting that it achieves acceptable
the tests are important issues. reliability as well.
This study was undertaken specifically to
extend study of the SFSTs from the laboratory
setting to field use. The primary study
question was, “How accurate are officers’
arrest and release decisions when the SFSTs
are used by trained and experienced officers?”
Over a five-month period, officers from seven
Colorado law enforcement agencies who

HS 178 R2/06 2
 III. Study Design

This study was designed to: In a broader context, the terms Correct
(1) gather data to assign officers’ Releases and Incorrect Releases could be
decisions to the four cells of the decision extended to motorists who were stopped but
matrix illustrated in Figure 1, and to who were not asked to perform the SFSTs. In
(2) examine the accuracy of the SFST many of those cases, the release decisions
battery when used in the widely varying were correct, but it is likely that some of there
weather conditions of Colorado winter, spring, were impaired drivers who were released
and summer months. without ever being asked to perform the
Both the design and the execution of the SFSTs. Those individuals and those decisions
study focused on the integrity, completeness, are of interest and would be included in an
and standardization of the data. assessment of overall proficiency in DUI
It is important to note how the study detection and arrest. In fact, the entire
population was defined and how the sample of population of impaired drivers, only some of
subjects was drawn. Subjects were a subset of whom are detected and stopped, is of interest
the population of drivers who were detained in terms of traffic safety. In a validation study
by police officers during the study period. of SFSTs, however, the subjects were only
They were drivers, both those arrested and those drivers who were asked to perform the
those released, who were stopped by police tests.
officers during the study period and who were
requested to perform the SFSTs. The officers’
decisions about those drivers have been
analyzed in terms of correct decisions (Correct
Arrests and Correct Releases) and errors
(Incorrect Arrests and Incorrect Releases).

HS 178 R2/06 3
 VI. Summary and Discussion

In 1995, there is a sound base of citizen may conclude that driving after
scientific evidence to support the use of 0.10%, drinking is acceptable behavior on other
0.08%, and 0.05% BACs as presumptive and occasions.
per se alcohol limits for drivers. There also For a number of reasons, the difficulties
appears to be strong support for those statutes associated with traffic officers’ alcohol-
among citizens throughout broad (though not enforcement responsibilities typically are
all) segments of society. A clear-cut shift of underestimated. One reason is the misnomer
attitude over the past ten to fifteen years has “drunk driving,” which suggests that their
resulted in anti-drunk driving sentiments by duty is to apprehend “drunks” or obviously-
much of the driving population. In many intoxicated individuals. If that were indeed
social circles drinking-and-driving now is the sole definition of alcohol enforcement
unacceptable behavior. duties, the task would be fairly
Why then, in a largely pro-alcohol straightforward. In reality, the risks
enforcement climate, are there negative views associated with drinking and driving are not
of traffic officers’ related activities? Citizens limited to obviously-intoxicated drivers, nor
often seem to believe that enforcement is hit- are officers’ enforcement responsibilities
or-miss and that officers regularly fail to restricted to those drivers.
remove many, if not most, alcohol-impaired Traffic officers are responsible for
drivers from the roadway. Some also seem to removing alcohol-impaired drivers from the
believe that the activities at roadside are roadway, and the Colorado statute sets the
arbitrary and calculated to harass. Although criterion alcohol levels at 0.10% and 0.05%
the multifaceted social and individual BAC. In other jurisdictions the BAC limit is
variables that underlie this paradox of 0.08%, with additional lower levels for lesser
concurrent anti-enforcement sentiment and charges and specific driver groups.
anti-drunk driving sentiment are beyond the Enforcement problems arise in part from the
scope of this report, it is germane to consider fact that although the evidence clearly
one set of factors. At least part of this view of establishes that driving skills are impaired at
alcohol enforcement is attributable to a 0.10% BAC and lower, many, possibly even
general failure to recognize the importance of most, individuals who are willing to drive after
traffic officers’ duties, and to understand not drinking are not obviously intoxicated at those
only what their duties encompass but also the levels.
difficulty of their task. Leaving aside the problem of detecting
Legislators, regulatory agencies, alcohol impairment by the observation of
activities groups, and safety-conscious citizens driving behaviors, consider officers’ task once
alike sometimes appear to overlook the fact they stop vehicles and contact drivers at
that traffic officers are pivotal in the roadside. Working under widely-varying
deterrence of drunk driving. Unless officers conditions without special measurement
are able to detect and arrest impaired drivers, apparatus, they must decide within a few
those drivers will never enter the system of minutes whether a specific driver is impaired
sanctions and, therefore, the existence of by alcohol. Impaired drivers may or may not
enabling statutes and anti-drunk driving display atypical speech, appearance, or other
sentiment will be largely irrelevant to them. personal characteristics, but in either
Unfortunately, it is also true that the escape of circumstance the officers have no knowledge of
detection and arrest on multiple occasions any given driver’s sober appearance and
serves to reinforce the risky behavior. In behavior. The task is further complicated by
effect, if no accident and no arrest occur on one the tolerant drinker’s normal appearance even
or more occasions of drinking and driving, the at very high BACs.

HS 178 R2/06 4
 Are there signs and symptoms which are In an initial study of field sobriety tests
reliably associated with 0.05% and 0.10%? with 238 laboratory subjects, officers’ decisions
With what level of confidence can the officer overall were 76% correct (Burns and
arrest or release a driver? With a decision Moskowitz, 1977). Only 54% of their arrest
criterion that minimizes incorrect arrests, the decisions were correct, and only 8% of their
risk of releasing impaired drivers rises. On release decisions were incorrect. In a second
the other hand, a very strict decision criterion laboratory study, officers’ decisions overall
will decrease the number of impaired drivers were 81% correct, their arrest decisions were
who are released but at the risk of 68% correct, and 14% of their release decisions
unnecessarily detaining non-impaired drivers. were wrong (Tharp, Burns and Moskowitz,
Is one risk preferable to the other? These 1981). It is apparent that the arrest criterion
questions define the context of traffic officers’ was lower in the laboratory. The penalties for
alcohol enforcement activities and the mistakes in a laboratory setting are, of course,
background of the Colorado Validation Study fairly trivial compared to a real-world setting.
of the SFSTs. The lower criterion, together with lack of
experience with the tests, accounts for higher
rates of incorrect arrests and lower rates of
The records collected and analyzed during incorrect releases than found in this study. It
this study provide evidence that the SFSTs, is not surprising to find that officers in the
as used at roadside by trained and field require more certainty about arresting a
experienced law enforcement officers, are citizen and adopt a higher criterion with the
valid indices of the presence of alcohol. result that they err in the direction of
incorrect releases.
In summary, the data provide clear-cut
Records of all driver contacts, which findings about the use of SFSTs by officers in
resulted in administration of the SFSTs six Colorado communities. On a broader scale,
during the study period, were entered into the they provide partial and tentative answers to
analysis. Overall, for 234 cases confirmed by some important questions. It is hoped that
breath or blood tests, officers’ decisions to current data from a field setting will facilitate
arrest and release were 86% correct, and 93% court proceedings with drivers arrested on
of their arrest decisions were correct. DUI and DWAI charges. It is hoped, too, that
It was not unexpected to find that officers the content of this report will add to the
were almost twice as likely to release driving public’s understanding of roadside
incorrectly as to arrest incorrectly. enforcement activities, as well as to
Nonetheless, only 36% of the released drivers recognition of police officers’ critical role in
were at or above the statutory limit. traffic safety.
These findings obtained in the field with
officers experienced with the use of SFSTs can
be compared with findings from a laboratory
setting with officers recently trained with the
SFSTs. It should be kept in mind that the
current data are not fully comparable to data
from laboratory experiments, since there are
differences other than time-since-training and
laboratory vs. field. With that caution, the
comparisons are instructive.

HS 178 R2/06 5
 ATTACHMENT C

A FLORIDA VALIDATION
STUDY OF THE
STANDARDIZED FIELD
SOBRIETY TEST (S.F.S.T.)
BATTERY

Marcelline Burns, Ph.D.

Southern California Research Institute
Los Angeles, California

Teresa Dioquino, Sergeant

Pinellas County Sheriff’s Office
Largo, Florida

HS 178 R2/06
I. INTRODUCTION

During the years 1975-1981, a battery of field sobriety tests was developed under
funding by the National Highway Traffic Safety Administration (NHTSA), U.S.
Department of Transportation (Burns and Moskowitz, 1977; Tharp, Burns, and
Moskowitz, 1981). The tests include Walk-and-Turn (WAT), One-Leg Stand (OLS),
and Horizontal Gaze Nystagmus (HGN). NHTSA subsequently developed a
training curriculum for the three-test battery, and initiated training programs
nationwide. Traffic officers in all 50 states now have been trained to administer the
Standardized Field Sobriety Tests (SFSTs) to individuals suspected of impaired
driving and to score their performance of the tests.

At the time the SFSTs were developed, the statutory blood alcohol concentration
(BAC) for driving was 0.10% throughout the United States. The limit now has been
lowered in a number of states to 0.08% for the general driving population. “Zero
tolerance” is in effect in some jurisdictions for drivers under age 21, and commercial
drivers risk losing their licenses at a BAC of 0.04%. It is likely that additional
states will enact stricter statutory limits for driving. In light of these changes, a re-
examination of the battery was undertaken by McKnight et al. (1995). They
reported that the test battery is valid for detection of low BACs and that no other
measures or observations offer greater validity for BACs of 0.08% and higher.

The three tests have been incorporated into Drug Influence Evaluations (DIEs)
which are conducted by certified Drug Recognition Experts (DREs) whenever an
individual is suspected of being drug-impaired. As part of a DRE evaluation, the
SFSTs provide important evidence of drug impairment and contribute to the DRE’s
three-part opinion:

# Is the individual impaired by a drug or drugs?

# If yes, is the impairment drug-related?
# If yes, what category or categories of drug account for the impairment?

A study was conducted in Colorado to examine the validity of the SFSTs when used
by experienced officers in the field (Burns and Anderson, 1995). The design of the
study insured that roadside testing was limited to the three-test battery, and that
officers’ decisions were not influenced either by the driver’s performance of other
behavioral tests or by measurement of BAC with a preliminary breath tester (PBT).
The obtained data demonstrated that more than 90% of the officers’ decisions to
arrest drivers were confirmed by analysis of breath and blood specimens.

A recently-reported NHTSA-funded study was conducted by Anacapa Sciences, Inc.

in collaboration with the San Diego Police Department to examine the validity of
the SFSTs for both 0.08% and 0.04% (Stuster and Burns, 1997). Officers’ estimates
of whether a driver’s BAC was above or below 0.08% or 0.04% were found to be
more than 90% correct.

HS 178 R2/06 2
The Colorado and California studies provide relevant and current field data. The
validity of the tests when they are administered in the context of drug evaluations
was examined in a retrospective analysis of the records of the Phoenix DRE Unit
(Adler and Burns, 1994). It was found that a suspect’s performance of the tests
provides valid clues of drug impairment.

The study reported here was conducted in collaboration with the Pinellas County
Sheriff’s Office (PCSO) and expands the examination of the SFSTs to the State of
Florida. An overview of PCSO and the demographics for Pinellas County can be
found in Appendix I.

II. STUDY BACKGROUND AND RATIONALE

During the early years of SFST use by law enforcement, legal challenges were
relatively infrequent. For more than a decade now, however, defense counsel in
many jurisdictions has sought to prevent the admission of testimony about a
defendant’s performance of the three tests. The objections, which continue to be
persistent and vigorous in 1997, typically focus on test validity and reliability as
demonstrated in the original laboratory research. It is entirely appropriate to
inquire whether that early research to identify a best set of sobriety tests was
conducted with scientific rigor. Beyond that inquiry, however, the data, which were
obtained in a laboratory setting and now are more than twenty years old, are of
little interest. Certainly, they are only marginally relevant to current roadside use
of the tests. The questions which begs to be addressed in 1997 is whether the tests
are valid and reliable indices of the presence of alcohol when they are used at
roadside under present day traffic and law enforcement conditions.

Experience and confidence have a direct bearing on an officer’s skill with roadside
tests. In this regard, note that the officers who participated in the early SCRI
studies had been only recently and briefly (4 hrs) trained to administer the test
battery. There had been no time for them to use the tests in the field where they
might have developed confidence in decisions based on them. Nonetheless, their
decisions were 76% correct in the first study and 81% correct in the second study.

At this point in time, many traffic officers have had ten or more years’ experience
with the test battery and many report that they confidently rely on them. Since it
seems unlikely in the extreme that they would continue to rely on tests which
repeatedly lead to decision errors, it is a reasonable assumption that more often
than not their roadside decisions to arrest are supported by measured BACs.
Whether their decisions to release are correct is largely unknown since the released
driver’s BAC generally is not measured.

HS 178 R2/06 3
Traffic officers are charged with the detection and arrest of impaired drivers.
Although their roadside duties are central to roadway safety, recognition of alcohol-
impaired drivers can be difficult and is, therefore, subject to error. If officers are to
effectively meet this particular enforcement responsibility, they need to augment
their general observations of suspects with sensitive, accurate sobriety tests. The
tests not only aid in the removal of dangerously impaired drivers from the roadway,
they also protect the driver who is not alcohol or drug impaired from being
improperly detained. Thus, rigorous examinations of the SFSTs are important to
traffic safety.

V. RESULTS

The first record in the data base is for an arrest which occurred on June 1, 1997,
and the last record is dated September 4, 1997. During the study period, 379
records were submitted for the study. Figure 3 graphs the total number of records
by month. As expected, the initial activities generated enthusiasm among
participants, and the largest number of citizen contacts occurred during the first
project month. Although available time of participating officers was affected during
July and August by scheduled training days and vacations, and although it typically
is difficult to sustain the initial high interest level, the actual decline in arrests over
the extended project period was not large. The final month is not comparable, since
data collection extended only a few days into September.

HS 178 R2/06 4
A. Total Sample and Measured BACs

Table 3 summarizes the disposition of 379 records obtained during this study. As
can be seen in the table and in Figure 4, the BACs of 256 drivers were measured.
Thus, BACs are available for 81.8% of the 313 cases entered into an analysis of
officers’ decisions. Evidential testing at the booking facility accounts for 210 of the
BACs. Forty-six were obtained with a Preliminary Breath Testing (PBT) device. A
log of all cases appears in Appendix IV.

VI. SUMMARY AND DISCUSSION

Legislators have lowered the limits for alcohol levels in drivers from 0.15%, which
was the very early standard, to 0.10% or 0.08%. The lower statutory limits are
soundly based in data from scientific experiments and form epidemiology and are
an important step toward safer roadways. Whether their full potential for reducing
alcohol-involved crashes can be reached, however, depends on effective enforcement.
Failure to enforce a statute, whatever the reason for the failure, weakens that
statute and may actually render it counterproductive to some degree.

Traffic officers are the first link in the series of events that brings a DUI driver into
the criminal justice system. Unless officers are able to detect and arrest impaired
drivers, those drivers will not experience the sanctions which are intended to deter
impaired driving. Although there are many aspects to effective DUI enforcement,
certainly it is crucial for officers to be proficient in assessing the alcohol impairment
of drivers they detain at roadside.

As an aid to their roadside decisions, officers rely upon a battery of tests, the
SFSTs, to augment their general observations of a driver. At this point in time, no
other tests have been shown to better discriminate between impaired and
unimpaired drivers. Nonetheless, the battery, and in particular Horizontal Gaze
Nystagmus, frequently is attached vigorously during court proceedings. Thus, the
examination of officers’ decisions, based on the SFSTs, is of considerable interest.

If it can be shown that officers’ reliance on the tests is misplaced, causing them
frequently to err, then the officers, the courts, and the driving public need to be
aware that the tests are not valid and that DUI laws are not bing properly enforced.
If, on the other hand, it can be shown that officer typically make correct decisions,
based on the SFSTs, perhaps the legal controversy that has centered on them for
more than a decade can be diffused and court time can be devoted to more
substantive issues.

HS 178 R2/06 5
The data obtained during this study demonstrate that 95% of the officers’ decisions
to arrest drivers were correct decisions. Furthermore, 82% of their decisions to
release drivers were correct. It is concluded that the SFSTs not only aid police
officers in meeting their responsibility to remove alcohol-impaired drivers from the
roadway, they also protect the rights of the unimpaired driver. These data validate
the SFSTs as used in the State of Florida by Pinellas County Sheriff’s deputies who
have been trained under NHTSA guidelines. SFST validity now has been
demonstrated in Florida, California (1997) and Colorado (1995). There appears to
be little basis for continuing legal challenge.

References

Burns, M. (1990) Development and Pilot Test of a Computer Data Base of Drug
Evaluations of Impaired Drivers. No. 90M079578301D.

Burns, M. and Adler, E.V. (1994) Drug Recognition Expert (DRE) Validation Study.
Final Report, E0072023, Governor’s Office of Highway Safety, State of
Arizona

Burns, M. and Anderson, E. (1995) A Colorado Validation Study of the

Standardized Field Sobriety Test (SFST) Battery. Final Report, Colorado
Department of Transportation.

Burns, M. and Moskowitz, H. (1977) Psychophysical Tests for DWI Arrest. Final
Report, DOT-HS-802 424, NHTSA, U.S. Department of Transportation.

McKnight, A.J., Langston, E.A., Lange, J.E., and McKnight, A.S. (1995)
Development of Standardized Field Sobriety Test for Lower BAC Limits
(SFST-LL), Final Report, DTNH22-92-07000, NHTSA, U.S. Dept. of
Transportation.

Stuster, J. and Burns, M. (1977) Validation of the Standardized Field Sobriety Test
Battery at BACs Below 0.10 Percent. Final Report, DTNH22-95-C-05192,
NHTSA, U.S. Dept. of Transportation.

Tharp, V., Burns, M. and Moskowitz, H. (1981). Development and Field Test of
Psychophysical Tests for DWI Arrests. Final Report, DOT-HS-805-864,
NHTSA, U.S. Dept. of Transportation.

HS 178 R2/06 6
VALIDATION OF THE STANDARDIZED F IELD SOBRIETY TEST
B ATTERY AT BAC S B ELOW 0.10 P ERCENT

F INAL REPORT

Submitted to:
U.S. DEPARTMENT OF T RANSPORTATION
N ATIONAL HIGHWAY T RAFFIC S AFETY A DMINISTRATION

Jack Stuster
Marcelline Burns

August 1998

A NACAPA SCIENCES , INC .

P.O. Box 519
Santa Barbara, California 93102
[Blank Page]

-- ii--
 Technical Report Documentation Page

1. Report No. 2. Government Accession No. 3. Recipient’s Catalog No.

DOT HS ?
4. Title and Subtitle 5. Report Date
Validation of the Standardized Field Sobriety Test Battery at August 1998
BACs Below 0.10 Percent
7. Author(s) 6. Performing Organization Code: n/a
Jack W. Stuster, PhD, CPE, and Marcelline Burns, PhD 8. Performing Organization Report No. n/a

9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Anacapa Sciences, Inc.
P.O. Box 519 11. Contract or Grant No.
Santa Barbara, CA 93102 DTNH22-95-C-05192
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
National Highway Traffic Safety Administration Final Report
400 Seventh Street, SW
Washington, D.C. 20590
15. Supplemental Notes 14. Sponsoring Agency Code
James F. Frank, PhD was the Contracting Officer’s Technical
Represenative (COTR) for this project.
16. Abstract
This study evaluated the accuracy of the Standardized Field Sobriety Test (SFST) Battery to assist officers in
making arrest decisions for DWI at blood alcohol concentrations (BACs) below 0.10 percent. NHTSA’s SFST battery
was validated at 0.10 percent BAC in 1981. The trend to reduce statutory DWI limits to 0.08 percent BAC prompted
this research project.
The research was composed of several project tasks, including planning, site-selection, training, data entry, and
data analysis, in addition to the actual conduct of a major field study. The City of San Diego, California, was selected
as the site. Seven officers of the San Diego Police Department’s alcohol enforcement unit were trained in the
administration and modified scoring of NHTSA’s SFST battery (i.e., Horizontal Gaze Nystagmus-HGN, Walk and Turn,
and One Leg Stand). SFST scoring was adjusted: the observation of four HGN clues indicated a BAC ≥0.08 percent
(rather than four clues indicating a BAC ≥0.10 percent), and the observation of two HGN clues indicated a BAC
≥0.04 percent. During routine patrols, the participating officers followed study procedures in administering SFSTs
and completing a data collection form for each test administered. The officers’ final step in each case was the
administration of an evidentiary breath alcohol test.
Data analysis found the SFSTs to be extremely accurate in discriminating between BACs above and below 0.08
percent. The mean estimated and measured BACs of the 297 motorists tested were 0.117 and 0.122, respectively;
the difference between the means (0.005 percent BAC) is very small and operationally irrelevant. Further, analyses
found the HGN test to be the most predictive of the three components of the SFST battery (r=0.65), however a
higher correlation was obtained when the results of all three tests were combined (r=0.69).
Decision analyses found that officers’ estimates of whether a motorist’s BAC was above or below 0.08 or 0.04
percent were extremely accurate. Estimates at the 0.08 level were accurate in 91 percent of the cases, or as high as
94 percent if explanations for some of the false positives are accepted. Officers’ estimates of whether a motorist’s
BAC was above 0.04 percent but lower than 0.08 percent were accurate in 94 percent of the decisions to arrest and
in 80 percent of cases overall. Also, the officers and prosecutors who were interviewed about the SFSTs found the
test battery to be acceptable for field use to establish probable cause for DWI arrest.
The results of this study provide clear evidence of the validity of the Standardized Field Sobriety Test Battery to
discriminate at 0.08 percent BAC, using a slightly modified scoring procedure. Further, study results strongly
suggest that the SFSTs also accurately discriminate at 0.04 percent BAC.
17. Key Words 18. Distribution Statement
DWI, DUI, detection, field sobriety, performance tests,
alcohol, law enforcement, highway safety
19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified

--iii--
[Blank Page]

-- iv --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

E XECUTIVE SUMMARY
This report documents the research activities and presents the results of a
study conducted for the National Highway Traffic Safety Administration (NHTSA)
to evaluate the accuracy of the Standardized Field Sobriety Test (SFST) Battery to
assist officers in making arrest decisions and to discriminate blood alcohol
concentrations (BACs) below 0.10 percent. NHTSA’s SFST battery was validated at
0.10 percent BAC in 1981. The trend to reduce statutory DWI limits to 0.08 percent
BAC prompted this research project.

DESCRIPTION OF THE RESEARCH

The research was composed of several project tasks, including planning, site-
selection, training, data entry, and data analysis, in addition to the actual conduct of
a major field study. The City of San Diego, California, was selected as the site of the
field study. Seven officers of the San Diego Police Department’s alcohol enforcement
unit were trained in the administration and modified scoring of NHTSA’s SFST
battery (i.e., Horizontal Gaze Nystagmus, Walk and Turn, and One Leg Stand). SFST
scoring was changed slightly: the observation of four horizontal gaze nystagmus
(HGN) clues indicated a BAC ≥0.08 percent (rather than four clues indicating a BAC
≥0.10 percent), and the observation of two HGN clues indicated a BAC ≥0.04 percent.
During routine patrols, the participating officers followed study procedures in
administering SFSTs and completing a data collection form for each test
administered during the study period. The officers’ final step in each case was the
administration of an evidentiary breath alcohol test.

RESULTS
The participating officers completed a total of 298 data collection forms; only
one case was eliminated from analysis because the motorist refused all forms of
BAC testing. Data analysis found the SFSTs to be extremely accurate in
discriminating between BACs above and below 0.08 percent. The mean estimated
and measured BACs of the 297 motorists tested were 0.117 and 0.122, respectively;
the difference between the means (0.005 percent BAC) is very small and
operationally irrelevant. Further, analyses found the HGN test to be the most
predictive of the three components of the SFST battery (r=0.65), however a higher
correlation was obtained when the results of all three tests were combined (r=0.69).

The results of decision analyses provide clear indication of SFST accuracy.

Decision analyses found that officers’ estimates of whether a motorist’s BAC was
above or below 0.08 or 0.04 percent were extremely accurate. Estimates at the 0.08
level were accurate in 91 percent of the cases, or as high as 94 percent if explanations
for some of the false positives are accepted. Officers’ estimates of whether a
motorist’s BAC was above 0.04 but under 0.08 were accurate in 94 percent of the
decisions to arrest and in 80 percent of the relevant cases, overall.

-- v --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

Finally, the officers and prosecutors who were interviewed about the SFSTs
found the test battery to be fully acceptable for field use to establish probable cause
for DWI arrest.

IMPLICATIONS
The results of this study provide clear evidence of the validity of the
Standardized Field Sobriety Test Battery to discriminate above or below 0.08 percent
BAC, using a slightly modified scoring procedure. Further, study results strongly
suggest that the SFSTs also accurately discriminate above or below 0.04 percent BAC.

-- vi --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

TABLE OF C ONTENTS

Page
INTRODUCTION ........................................................................................................................ 1
Background...................................................................................................................... 1

THE RESEARCH ....................................................................................................................... 5

Task 1: Refined Work Plan........................................................................................... 5
Task 2: Specified SFSTs and Revised Procedures..................................................... 5
Task 3: Selected and Recruited Law Enforcement Agency..................................... 5
Task 4: Conducted the Field Validation Study......................................................... 9
Tasks 5 and 6: Analyzed Data and Prepared Final Report...................................... 14

RESULTS ................................................................................................................................. 15
Evaluation of SFST Accuracy....................................................................................... 15
Comparison of Means.............................................................................................. 15
Correlation Analyses................................................................................................ 17
Decision Analyses..................................................................................................... 18
Evaluation of SFST Acceptability................................................................................ 23

IMPLICATIONS .......................................................................................................................... 25
How Accurately Do the Tests Discriminate?............................................................ 25
Which of the Components of the SFST Battery Is/Are the Best Predictors?..... 26
How Reliable, or Consistent, Are the Tests?............................................................. 26
Are the Tests Usable By Officers? Are They Accepted by Officers and
Prosecutors?..................................................................................................................... 27
Conclusions...................................................................................................................... 28

REFERENCES ............................................................................................................................. 29

A PPENDIX A: STANDARDIZED FIELD SOBRIETY TESTING..................................................... 31

-- vii --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

LIST OF TABLES

Table Page

1 COMPONENTS OF THE FIELD EXPERIMENT PLAN......................................................... 9

2 ESTIMATED AND M EASURED BAC BY A GE CATEGORY ............................................ 15
3 ESTIMATED AND M EASURED BAC BY GENDER ......................................................... 16
4 ESTIMATED AND M EASURED BAC BY DISPOSITION, A GE CATEGORY, AND GENDER 16
5 CORRELATIONS OF SFST S CORES TO ESTIMATED AND M EASURED BAC.................. 17
6 S UMMARY OF FALSE POSITIVES .................................................................................. 19
7 S UMMARY OF FALSE N EGATIVES ............................................................................... 20

F IGURES

Figure Page

1 Fatality rates per million miles traveled in the U.S....................................... 2

2 Sequence of major project tasks.......................................................................... 5
3 Data collection form used in the validation study......................................... 12
4 Decision matrix at 0.08 percent BAC.................................................................. 18
5 Decision matrices at 0.08 percent BAC for each component test................. 21
6 Venn diagram of 188 cases ≥ 0.08 percent BAC................................................ 22
7 Decision matrix at 0.04 percent BAC.................................................................. 23

-- viii --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

ACKNOWLEDGMENTS

The National Highway Traffic Safety Administration and Anacapa Sciences,

Inc., are grateful to the managers and officers of the San Diego Police Department for
their cooperation during the performance of this research project. Special thanks to
Officer Doug English, project liaison, and Sergeant Bill Snyder, supervisor of the
Traffic Division’s alcohol-enforcement unit.

-- ix --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

-- x --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

INTRODUCTION

Beginning in 1975, the National Highway Traffic Safety Administration

(NHTSA) sponsored research that led to the development of standardized
methods for police officers to use when evaluating motorists who are suspected of
Driving While Impaired (DWI).1 Beginning in 1981, law enforcement officers have
used NHTSA’s Standardized Field Sobriety Test (SFST) battery to help determine
whether motorists who are suspected of DWI have blood alcohol concentrations
(BACs) greater than 0.10 percent. Since that time, many states have implemented
laws that define DWI at BACs below 0.10. This report presents the results of
research performed to systematically evaluate the accuracy of NHTSA’s SFST
battery to discriminate above or below 0.08 percent and above or below 0.04 percent
blood alcohol concentration.

The report is presented in four sections. This brief Introduction presents the
objectives of the research, provides a summary of the relevant traffic safety issues,
and discusses the historical context of the study. The second section of the report
describes the research tasks that were performed. The third section presents the
results of the study. The final section of the report discusses the implications of the
study results.

BACKGROUND
Nearly 1.4 million people have died in traffic crashes in the United States
since 1966, the year of the National Traffic and Motor Vehicle Safety Act (which
led to the creation of NHTSA in 1970). During the late 1960s and early 1970s more
than 50,000 people lost their lives each year on our nation’s public roads; more
than half of the motorists killed had been drinking. Traffic safety has improved
considerably since that time: the annual death toll has declined to about 40,000,
even though the numbers of drivers, vehicles, and miles driven all have greatly
increased. The dramatic improvements in traffic safety are reflected in the change
in fatality rate per 100 million vehicle miles traveled: The fatality rate fell from 5.5
in 1966 to 1.7 in 1996 (FARS--Fatal Analysis Reporting System--96), a 69 percent
improvement. Figure 1 illustrates this important trend. When miles traveled are
considered, the likelihood of being killed in traffic in 1966 was more than three
times what it is today.

Despite the significant improvements in traffic safety during the past 17

years, an average of more than 115 people still die each day from motor vehicle
crashes in the United States. It is estimated that 41 percent of the drivers who die
in crashes have been drinking.

________________________________
1
Various terms are used throughout the United States for offenses involving drinking and driving. In
this report, Driving While Impaired (DWI) is used to refer to all occurrences of driving at or above
the legal blood alcohol concentratiion (BAC) limit of a jurisdiction.

-- 1 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

An emphasis on DWI enforcement since 1980 has been a factor in the

significant improvement in traffic safety, as represented by declining fatal and
alcohol-involved crash rates. NHTSA-sponsored research contributed substantially
to the improved condition, in part, by providing patrol officers with useful and
scientifically valid information and training materials concerning the behaviors
that are most predictive of impairment. In particular, NHTSA sponsored research
that led to the development of a DWI detection guide that listed 20 driving cues and
the probabilities that a driver exhibiting a cue would have a BAC of at least 0.10
percent (Harris et al., 1980; Harris, 1980). A similar study was conducted recently that
identified 24 driving cues that are predictive of DWI at the 0.08 level (Stuster, 1997).
NHTSA also sponsored research that led to the development of a motorcycle DWI
detection guide (Stuster, 1993). NHTSA’s DWI training materials, based on the
results of these studies, have exposed the current generation of law enforcement
officers in the U.S. to information critical to DWI enforcement by providing a
systematic, scientifically valid, and defensible approach to on-the-road DWI
detection.

6
Fatality Rate per Million Vehicle Miles Travelled

4 DWI Detection
Training Begins
SFSTs Are
Introduced

0
1967

1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980

1982
1983
1984
1985
1986

1989
1990
1991
1992
1993
1966

1968

1981

1987
1988

1994
1995

Figure 1. Fatality rates per million miles traveled in the U.S.

At the same time NHTSA was providing patrol officers with information
concerning the driving behaviors that are the most predictive of impairment, the
agency also sponsored research that led to the development of a standardized battery

-- 2 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

of tests for officers to administer to assess driver impairment after an enforcement

stop has been made. Drs. Marcelline Burns and Herbert Moskowitz conducted
laboratory evaluations of several of the tests that were most frequently-used by law
enforcement officers at the time (Burns and Moskowitz, 1977). In addition to a
variety of customary roadside tests (e.g., finger-to-nose, maze tracing, backward
counting), the researchers evaluated measures of an autonomic reaction to central
nervous system depressants, known as horizontal gaze nystagmus. Horizontal gaze
nystagmus (HGN) is an involuntary jerking of the eye that occurs naturally as the
eyes gaze to the side. Aschan (1958) described studies that linked various forms of
nystagmus to BAC, and Wilkinson, Kime, and Purnell (1974) reported consistent
changes in horizontal gaze nystagmus with increasing doses of alcohol. At the time
Burns and Moskowitz were conducting their seminal research for NHTSA,
horizontal gaze nystagmus recently had been found to reliably predict BACs in a
study conducted in Finland (Pentilla, Tenhu, and Kataja, 1974). Further, Lehti (1976)
had just calculated a strong correlation between BAC and the onset of nystagmus.
All of the field sobriety tests evaluated by Burns and Moskowitz were found
to be sensitive to BAC in varying degrees, at least under laboratory conditions. In
addition, all of the tests showed a consistent increase in correlations with increasing
BACs. Statistical analyses found the horizontal gaze nystagmus test to be the most
predictive of the individual measures. However, the combined scores of three of the
tests (One-Leg Stand, Walk-and-Turn, and Horizontal Gaze Nystagmus) provided a
slightly higher correlation than the horizontal gaze nystagmus test by itself. The
combined score correctly discriminated between BACs below or above 0.10 in 83
percent of the subjects tested in the original study (Burns and Moskowitz, 1977).
NHTSA immediately sponsored a subsequent study to standardize the test
administration and scoring procedures and conduct further laboratory and field
evaluations of the new battery of three tests. The researchers found that police
officers tended to increase their arrest rates and were more effective in estimating
the BACs of stopped drivers after they had been trained in the administration and
scoring of the Standardized Field Sobriety Test battery. The results of this important
study were documented in meticulous detail in the technical report, Development
and Field Test of Psychophysical Tests for DWI Arrest (Tharp, Burns, and
Moskowitz, 1981). That report has been cited throughout the U.S. to establish the
scientific validity of the SFST battery and to support officers’ testimony in court.
NHTSA’s SFST battery is described in Appendix A.
During the past 16 years, NHTSA’s SFSTs largely have replaced the
unvalidated performance tests of unknown merit that once were the patrol officer’s
only tools in helping to make post-stop DWI arrest decisions. Regional and local
preferences for other performance tests still exist, even though some of the tests
have not been validated. Despite regional differences in what tests are used to assist
officers in making DWI arrest decisions, NHTSA’s SFSTs presently are used in all 50
states. NHTSA’s SFSTs have become the standard pre-arrest procedures for
evaluating DWI in most law enforcement agencies.

-- 3 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

The horizontal gaze nystagmus (HGN) test is considered by many law

enforcement officers to be a foolproof technique (sometimes called a “silver bullet”)
that provides indisputable evidence of alcohol in a motorist’s system. The normal
variation in human physical and cognitive capabilities, and the effects of alcohol
tolerance, result in uncertainties when arrest decisions are made exclusively on the
basis of performance tests. These uncertainties have resulted in large proportions of
DWI suspects being released rather than detained and transported to another
location for evidentiary chemical testing. This is important because experienced
drinkers often can perform physical and cognitive tests acceptably, with a BAC
greater than 0.10 percent. However, most experienced drinkers cannot conceal the
physiological effects of alcohol from an officer skilled in HGN administration. This
is because horizontal gaze nystagmus is an involuntary reaction over which an
individual has absolutely no control.

-- 4 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

THE RESEARCH

This section provides a detailed description of all tasks performed during the
field validation of the Standardized Field Sobriety Test Battery for use at 0.08 percent
BAC. The technical approach to the research involved the performance of six major
project tasks, as summarized in Figure 2 and described in the following pages.

Task 3:
Task 2:
Task 1: Selected/Recruited LE Agency,
Specified SFSTs and
Refined Work Plan Revised Training Program,
Revised Procedures
and Conducted Training

Task 4: Task 5: Task 6:

Conducted Field Entered and Analyzed Prepared Final
Study Data Report

Figure 2. Sequence of major project tasks.

TASK 1: REFINED WORK PLAN

The objectives of the first project task were to meet with the Contracting
Officer’s Technical Representative (COTR) and other NHTSA SFST experts to
discuss the project and to refine the proposed Work Plan based on those discussions.
The project kick-off meeting was held at NHTSA headquarters on 24 October 1995.
Substantive discussions with NHTSA personnel during and following the meeting
contributed to the development of the technical approach described here.

TASK 2: S PECIFIED SFST S AND REVISED PROCEDURES

Based on the widespread use and acceptance of NHTSA’s Standardized Field
Sobriety Test (SFST) Battery, validated at 0.10 percent BAC, NHTSA sponsored the
current study to evaluate the SFSTs at lower BACs. The only modifications to be
made to the SFSTs would be: 1) for officers to use the exhibition of four clues as an
indication of BACs at the 0.08 level or greater (as officers presently are trained to use
four clues as an indicator of BACs at 0.10 percent or greater), and 2) for officers to use
the exhibition of two HGN clues as an indication of BACs greater than zero, but
below 0.08 percent.

TASK 3: S ELECTED AND RECRUITED LAW E NFORCEMENT AGENCY AND

C ONDUCTED TRAINING
This project task was composed of four subtasks, as described in the following
paragraphs.

-- 5 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

SUBTASK 3.1: IDENTIFIED SITE SELECTION CRITERIA

The site-selection criteria were:
• Candidate sites must employ lower legal BAC levels (0.08 for adults and zero
tolerance for youth under 21 years).

• Candidate sites must generate a sufficient number of traffic enforcement stops and
DWI arrests for accurate assessment of the tests’ reliability and validity.

• Participating officers must have received NHTSA-approved SFST training from a

certified instructor, possess at least one year of field experience administering
SFSTs, and receive refresher training from project staff.

• Managers and officers of the participating law enforcement agency must agree to
abide by the research procedures for the duration of the field study. For example,
officers may use only the SFST Battery (and no other tests) together with their
observations of the driver’s general appearance and speech to make their arrest
decisions; and, all test administrations must be recorded and submitted. Only
agencies that could assure an extremely high level of cooperation and commitment
would be recommended for participation.

• The site must have the capability of generating cases that represent the full range
of alcohol experience. For example, a city with a disproportionate number of
younger drivers might be more appropriate to ensure samples of sufficient size for
the younger age categories.

SUBTASK 3.2: IDENTIFIED CANDIDATE SITES AND A PPLIED SELECTION CRITERIA

Several factors constrained the site-selection process and limited the possible
candidates for participation in this study. First, at the time the project was
conducted, California, Oregon, and Utah were the only states that met both of the
BAC-related site-selection criteria, namely a 0.08 BAC limit for DWI and a zero
tolerance law for drivers under 21 years of age. Second, it was important to restrict
the data collection period, to the extent possible, because it was believed that an
extremely long data collection period might result in officers deviating from the
study procedures. Strict adherence to study procedures was considered essential to
ensuring the internal validity of the study.
The site-selection strategy adopted was to recruit a police department that
serves one large city--a city large enough to generate a sufficient number of SFST
administrations for statistical analysis by itself. A large city also was likely to have a
traffic division with a dedicated DWI unit composed of trained experts. Focusing on
traffic enforcement specialists would permit us to restrict participation in the study
to officers who already had received NHTSA-approved SFST training and had
additional field experience administering the test battery. Prior training in SFST
administration was an important site-selection and methodological issue.
In the study that validated the SFST battery in 1981, all officers of an agency
could participate, following training provided by the researchers. The procedure
followed during the original study was appropriate then because no other officers
(anywhere) had yet to receive the training. However, that procedure could not be
followed in the current study because thousands of officers have received SFST

-- 6 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

training since 1981. Only trained and experienced test administrators could be
permitted to participate in the current study to avoid confounding study results
with the effects of substantially different officer skill and experience levels in SFST
administration and scoring. Officers who are formally trained and experienced in
SFST administration tend to be concentrated in traffic enforcement and special DWI
units.
This site-selection strategy was judged to provide the best approach to achieve
the objectives of the current study, and the City of San Diego, California, was
identified as the leading candidate community when the site-selection criteria were
applied. The San Diego Police Department serves a resident population of more
than one million, with a much larger service population attributable to tourism and
several local military installations. The manner in which the San Diego Police
Department satisfied the site-selection criteria is outlined below.
Number of SFST Administrations
The San Diego Police Department maintains a traffic division composed of 50
officers, including ten officers and a sergeant who form the alcohol enforcement
unit. The alcohol enforcement unit deploys four or five officers on each night,
Wednesday through Sunday. The time necessary to complete the associated
paperwork usually limits each officer to a maximum of two DWI arrests each night.
This results in about 130 arrests by officers of the special unit during a four week
period. The other members of the traffic division, combined, make an additional 130
DWI arrests each month. San Diego Police Department officers do not hesitate to
arrest drivers for BACs below 0.08 percent if they exhibit any evidence of
impairment, even though low-BAC arrests usually are not prosecuted by the local
district attorney.
Demographic Considerations
The Work Plan discussed the importance of selecting a site that offers cases
for analysis that represent the full range of driver ages and BACs of interest. It was
believed that a younger, rather than an older, driver population would result in
more cases of zero tolerance violations and more SFST administrations overall. In
this regard, San Diego and the surrounding area is home to four major US Navy
bases and both the Navy and Marine Corps training centers. The area also is home
to three major universities and several smaller colleges and technical schools.
Willingness to Participate
Naturally, formal approval by senior managers is required before any law
enforcement agency can participate in a traffic safety study. Further, a manager’s
personal interest in a study that results in command emphasis concerning
participation greatly contributes to the success of a project because of the quasi-
military organizational structure of law enforcement agencies. That is, if managers
believe participation to be of value to an agency they will direct their officers to
follow the study procedures. In this regard, the commanding officer and other
senior managers of the San Diego Police Department expressed their considerable
interest in the study and directed their personnel to cooperate with the study team.

-- 7 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

Command emphasis is an important component to ensure adherence to

study procedures, but it is not sufficient; the participating officers also must be
committed to the study. The willingness of a law enforcement agency to participate
in a traffic safety study also can be measured, although subjectively, by the attitudes
of field officers when discussing the general and specific issues involved in the
study. The officers of the San Diego Police Department with whom we spoke about
the field validation expressed genuine interest in the study and eagerness to be
selected for participation.
Finally, the requirement for an agency to modify its established procedures to
accommodate special study procedures usually is somewhat negotiable in a traffic
safety study, but deviations from established study procedures were not negotiable
in this field validation. It was explained that police managers and all participating
officers must agree to abide by the study procedures to ensure the internal validity of
study results. This was an area for concern to the project team because the San Diego
Police Department’s established DWI procedures included administering three field
sobriety tests in addition to the three NHTSA SFSTs. A firm study requirement was
that no other tests be administered to subjects because they might influence an
officer’s BAC estimates; that is, all officer-estimates of BAC must be based
exclusively on results of the NHTSA SFST battery using the slightly modified
scoring system. In this regard, San Diego police managers inquired with their district
attorney and DWI supervisors, those who might object to the restriction, and found
no opposition. In fact, it was mentioned that restricting sobriety testing to the three
SFSTs would help streamline the procedures for everyone.
Prior SFST Training
All members of the San Diego Police Department’s special alcohol-enforce-
ment unit previously had received SFST training that was administered according
to NHTSA-approved procedures and curriculum by certified DWI instructors.
Although approximately half of the other members of the Traffic Division also had
received SFST training, it was determined that the alcohol-enforcement unit would
generate a sufficient number of SFST administrations for statistical analysis. All of
the participating officers would receive a four-hour refresher training course prior
to beginning the field study.
SUBTASK 3.3: RECRUITED LAW ENFORCEMENT A GENCY TO PARTICIPATE IN THE STUDY
NHTSA reviewed the site recommendations and approved San Diego as the
site for the field study. Further discussions were held with managers and officers of
the San Diego Police Department and a Memorandum of Agreement was signed
that specified all study procedures and requirements.
SUBTASK 3.4: DEVELOPED SFST TRAINING PROGRAM
The experimental requirement that all participating officers be both trained
and experienced in SFST administration eliminated the need to develop a special
training program for this study. It was considered essential that the existing,
NHTSA-approved SFST training program remain the training standard for the field
evaluation. Because all participating officers already had received NHTSA-approved

-- 8 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

SFST training, only a refresher program would be required. A four-hour refresher-

training program was developed, based on the (October 1995) NHTSA curriculum.
The purposes of the refresher training were to instruct the officers concerning the
modified scoring system and obtain confirmation that all participants were
administering and scoring the SFST battery correctly before beginning the field
study.

TASK 4: C ONDUCTED THE F IELD VALIDATION S TUDY

Systematic evaluation of the SFSTs to assist officers in making arrest
decisions at BACs below 0.10 percent, under field conditions, was the ultimate
objective of this research. Although existing tests were the subject of the evaluation,
the reasons for conducting the field study were the same as if the tests previously
had not been validated. First, it was necessary to determine the accuracy of the
modifications to test scoring, compared to actual BAC levels measured through
other means. For cases in which the driver was arrested for DWI, correspondence
would be assessed between scored performance on the SFSTs and BAC, as
determined by breath test (blood and urine tests were discouraged but used if
subjects refused to comply with breath testing). For cases in which a subject was
administered SFSTs but then released on the basis of low estimated BAC, hand-held
breath testing devices were used to establish actual BAC. The second purpose of the
evaluation was to identify problems with test application in the field, which might
include test administration, scoring procedures, or other factors that might affect the
use of the tests by law enforcement personnel. Third, the courts’ acceptance of
evidence gathered using the slightly revised scoring procedures in the field
evaluation would be assessed.
SUBTASK 4.1: PREPARED FIELD EXPERIMENT PLAN
A Field Experiment Plan was developed and approved by NHTSA to guide
the conduct of the field study. The plan included the seven components depicted in
Table 1 and discussed below.
T ABLE 1
COMPONENTS OF THE FIELD EXPERIMENT PLAN

Component 1: Subjects
Component 2: Independent Variables
Component 3: Criterion Measures
Component 4: Materials
Component 5: Procedures
Component 6: Controls
Component 7: Data Analyses

Components 1 and 2: Subjects and Independent Variables

The primary independent variable of interest, BAC, was inextricably linked to
the subjects in this study. Specifically, the experiment plan focused on obtaining data
from adult motorists who were suspected of exceeding the legal limit of 0.08 percent
BAC and youths under 21 who were suspected of exceeding the “zero-tolerance”

-- 9 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

legal limit of 0.00. The accuracy of the SFSTs to discriminate at 0.08 and 0.04 percent
BAC could not be assessed without data from individuals who had BACs over and
under these values. Therefore, it was important to obtain BAC estimates from
individuals who had both passed and failed the standardized field sobriety tests.
Component 3: Criterion Measures
The only appropriate criterion measure to assess the accuracy of SFSTs is
BAC. Measures of impairment are irrelevant because performance of the SFSTs
must be correlated with BAC level, rather than driving performance. BAC provides
an objective and reliable measure that states have recognized as presumptive
and/or per se evidence of impairment, depending on the statute. To obtain these
criterion measures, it was determined that all drivers who were administered the
SFST Battery must be tested for BAC, regardless of the results of the SFSTs. In other
words, it would be essential to test the individuals who were judged to have BACs
below the relevant statutory level and who subsequently would be released.
Participating officers were instructed concerning the importance of obtaining BAC
data for all subjects, in order to calculate the accuracy of the tests.
All police officers participating in the study were equipped with NHTSA-
approved, portable breath testing devices to assess the BACs of all drivers who were
administered the SFSTs, including those who were released without arrest. Further,
arrested subjects were tested both in the field with a portable device and at the
booking site. The use of passive alcohol sensors (PAS) during the study was not
permitted.
Component 4: Materials
Only the existing SFSTs were to be administered, which require no
equipment. A pen, pencil, or small flash light frequently are used by officers as a
stimulus or target for the HGN test, but a finger can be used with equal effectiveness.
The data collection form used in the study is presented as Figure 3. The data
collection form was extremely important in this study for several reasons. As is the
case in most field studies, the form must be as simple to complete as possible to
minimize the workload of participating officers. In the present case, it also was
important for the form to be designed to guide the officer in the administration of
the SFSTs, to facilitate standardization and systematic scoring of the tests. In
addition, the form designed for this study had to both encourage and provide
assurances that officers had followed the study procedures. Most important, it was
essential that officers would conduct a breath test and record actual subject BAC as
the final step of the process; that is, actual BACs were to be entered on the form only
after BAC estimates based on SFST performance had been recorded. Hand-held
breath testing devices with digital displays were used for this purpose.
Component 5: Procedures
The sixth component of the field experiment plan was the specification of
procedures to be used for administering the tests and obtaining independent
measures of BAC. The procedures to be followed by participating officers were listed

-- 10 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

as a series of six numbered steps on the data collection form that was used in the
field study. The study procedures were to be followed whenever a participating
officer suspected an adult driver of being alcohol impaired or a youth under 21 of
having a BAC greater than zero. In practice, officers administered the SFSTs to all
motorists who exhibited any objective behavior or other cue associated with having
consumed alcohol, even if impairment was not evident. A breath, blood, or urine
test was administered to all motorists who performed the SFSTs, but only after the
officer had made an arrest/no arrest decision based on the officer’s scoring of the
driver’s SFST performance, and recorded a BAC estimate. The data collection form
structured the procedure by presenting all officer actions as a series of numbered
steps. Requiring officers to record the time of BAC estimates and BAC tests ensured
that officers’ estimates were not influenced by the results of the chemical tests.
Completed data collection forms were sent to Anacapa Sciences on a weekly basis for
data entry.
In some states, such as California, officers have the right to administer a
breath test to a driver who has exhibited any objective sign of alcohol-consumption.
Compliance is mandatory if the officer can articulate a reasonable suspicion of the
motorist having consumed alcohol (such as the odor of an alcoholic beverage).
SFSTs were administered only to drivers who exhibited some objective DWI cue,
thus, no problems were experienced in obtaining BAC data, even from subjects
whose SFST performance was acceptable. The field breath test was conducted as the
final step after the SFST procedure was completed, which is the de facto procedure
followed by most officers who are equipped with field breath testing devices.
To further ensure compliance with study procedures, the participating law
enforcement officers signed a statement affirming that they would abide by the
established study procedures. In addition, project staff monitored the data collection
effort, periodically riding along with participating officers to ensure that study
procedures were being followed.
Component 6: Controls
Extraneous variables that could affect the outcome of the study must be
controlled to the extent possible. The controls that were implemented to ensure the
validity of study results have been discussed in this section, including systematic
procedures and the use of only trained and experienced officers.
Component 7: Data Analyses
The data analysis plan was designed to answer the following research
questions.
• How accurately do the tests discriminate between subjects who are above or below
0.08 and 0.04 percent BACs?

• Which of the components of the SFST battery is/are the best predictor(s) of BAC?

• How reliable, or consistent, are the tests?

• Are the tests usable by police officers? Are they readily accepted by officers and
prosecutors?

-- 11 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

NHTSA/ANACAPA SFST VALIDATION DATA FORM

Adult Male
Officer ID:______________ Driver:
Under 21 Female
Age:_______
Month_____Day_____1996 Time of Stop:_______hr _______min
FIELD SOBRIETY TESTS ADMINISTERED √
1. HORIZONTAL GAZE NYSTAGMUS TEST
Clues
Right Eye Left Eye
Lack of smooth pursuit
Nystagmus at maximum deviation
Nystagmus onset before 45 degrees
Clues + =
Total HGN Clues (6 clues maximum)
4 or more ≥ 0.08 / 2 or more ≥ 0.04
Clues
2. ONE LEG STAND TEST (seconds) 0-10 11-20 21-30
Sways while balancing
Uses arms for balance
Hops to maintain balance
Puts foot down
Cannot perform test (4 clues -- maximum)
Total One Leg Stand Clues
2 or more ≥ 0.08
Clues
3. WALK AND TURN TEST
Loses balance while listening to instructions
Starts before instructions are finished
1st 9 2nd 9
Stops while walking
Does not touch heel to toe
Steps off the line
Raises arms for balance
Incorrect number of steps
Trouble with turn (explain)___________________
Cannot perform the test (8 clues -- maximum)
Total Walk and Turn Clues
2 or more ≥ 0.08

4. ESTIMATE OF BAC BASED ON SFSTS:

Time of estimation _______hr _______min

5. SUBJECT BAC Refused

PBT Time of PBT test _______hr _______min

Other Time of other test _______hr _______min

Breath Blood Urine
6. DISPOSITION: Warning Citation DUI Arrest

Figure 3. Data collection form used in the validation study.

-- 12 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

SUBTASK 4.2. TRAINED OFFICERS IN THE U SE OF THE SFSTS

Dr. Marcelline Burns, one of the investigators who developed the SFST
battery, developed and conducted the refresher training for the participating
officers. Dr. Burns’ research and training experience in this field ensured that
officers received effective and credible refresher instruction. Dr. Burns was assisted
in the training session by the project director and NHTSA COTR.

SUBTASK 4.3. IMPLEMENTED EXPERIMENTAL DESIGN AND COLLECTED DATA

Implementation of the experiment design began immediately following the
completion of officer refresher training on 23 May 1996 and continued through 9
November. Specific study procedures were:
• Only officers who were members of the San Diego Police Department’s alcohol-
enforcement unit and who received NHTSA-approved SFST training participated
directly in the study. Dr. Marcelline Burns provided brief “refresher” training to all
participating officers to ensure a consistent and systematic approach to SFST
administration during the study.

• Upon commencement of the study period, participating officers used only the SFST
Battery (i.e., Horizontal Gaze Nystagmus, Walk and Turn, One Leg Stand) together
with their observations of a driver’s general appearance and speech, to establish
inferences about a subject for whom there was reasonable suspicion of driving while
impaired. In other words, no tests other than the three SFSTs were performed.

• Participating officers performed the administration steps in the sequence specified on

the data collection form; that is, they,

1. Administered the Horizontal Gaze Nystagmus test and recorded results.

2. Administered the One Leg Stand test and recorded results.

3. Administered the Walk and Turn test and recorded results.

4. Used the scoring systems that were printed on the data collection form (by
counting test “clues”) to estimate the subject’s BAC. Recorded their estimate of
the subject’s BAC based on SFST performance, together with their observations of
the subject’s general appearance and speech. Also, they recorded the time when
their estimate was made.

5. Checked the box that indicated the disposition of the stop: Warning, Citation, or
Arrest.

6. Recorded the subject’s BAC obtained from a field breath test; or, checked the
appropriate box for other tests or responses. Blood and urine test results were
provided later; every effort was made to obtain a breath test result for a l l
subjects. Recorded the time when the BAC test was performed.

• Obtained a BAC for all subjects who were administered SFSTs as the final step in the
test administration procedure. BACs were obtained for all subjects tested including
those subjects who officers estimated, on the basis of SFST results, to have BACs
below the legal limit.

• Participating officers completed and submited a data collection form for each subject
tested during the study period; that is, all administrations of the SFST battery by

-- 13 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

participating officers were recorded on a data collection form and submitted for
analysis.

• All completed data collection forms were sent to Anacapa Sciences, Inc., for data
entry and analysis.

SUBTASK 4.4 CONDUCTED COURT AND POLICE INTERVIEWS

The final data collection task was the conduct of open-ended interviews with
participating police officers and prosecutors who were exposed to the new SFSTs
during DWI cases. The purposes of the interviews were to determine if the tests
were acceptable to the officers for use in the field and to the prosecutors for use of
test results in court.

TASKS 5 AND 6: ANALYZED DATA AND PREPARED F INAL REPORT

All data collection forms were returned to Anacapa Sciences, Inc., sequentially
numbered, and the contents entered into a computerized data base. Data analyses
were performed by the project director and Dr. Marcelline Burns. The results of
those analyses are presented in the following section of this report.

-- 14 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

RESULTS
This study was conducted to evaluate the accuracy of NHTSA’s Standardized
Field Sobriety Test Battery in assisting officers to make arrest decisions at BACs
above and below 0.08 percent under field conditions. A secondary objective of the
study was to evaluate the possibility that the test battery also could be used to assist
officers in making arrest decisions at BACs lower than 0.08 percent.
The seven participating officers from the San Diego Police Department’s
alcohol-enforcement unit completed a total of 298 data collection forms during the
study period; only one case was eliminated from analysis because the subject refused
to submit to any form of BAC testing. Officer compliance with study procedures and
motivation to participate in the study remained high throughout the data collection
period.
E VALUATION OF SFST ACCURACY
Three methods were used to evaluate the accuracy of the SFST battery to
discriminate at the BACs of interest: comparison of means, correlation analyses, and
decision analyses.
COMPARISON OF MEANS
Table 2 presents a summary of the estimated and measured BAC data by age
category. The table shows that 91.9 percent of the motorists tested were adults,
compared to 8.1 percent youth, defined as motorists under the age of 21 years. The
mean estimated and measured BACs of the younger motorists were approximately
0.035 lower than the BACs of the adults tested during the field study. The officers’
mean estimated BACs, however, were very close to the mean measured BACs for
both adults and youth; on average, the difference between officers’ estimates and the
actual BACs were only 0.005 percent for adults and 0.007 percent for youth.
T ABLE 2
ESTIMATED AND M EASURED BAC (%) BY A GE CATEGORY

Age Estimated Measured

Category Number Percent BAC (Mean) BAC (Mean)

Adults 273 91.9 0.120 0.125

Youth 24 8.1 0.083 0.090
Total 297 100.0 0.117 0.122

Table 3 presents a summary of the estimated and measured BAC data by gen-
der category. The table shows that 87.9 percent of the motorists tested were males,
compared to 12.1 percent females, with adults and youth combined. The mean esti-
mated BACs of the male and female motorists tested were identical (i.e., 0.117 per-
cent). Again, for both categories, the officers’ mean estimated BACs were very close
to the mean measured BACs; on average, the difference between officers’ estimates
and the actual BACs were only 0.004 percent for males and 0.012 percent for females.

-- 15 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

T ABLE 3
ESTIMATED AND M EASURED BAC (%) BY GENDER

Estimated Measured
Gender Number Percent BAC (Mean) BAC (Mean)

Male 261 87.9 0.117 0.121

Female 36 12.1 0.117 0.129
Total 297 100.0 0.117 0.122

Table 4 presents a more detailed accounting of the estimated and measured

BAC data by age and gender category, and by the disposition of the enforcement stop.
In addition, the table shows that 73 percent of all motorists who were tested during
the field study were arrested for DWI based on SFST performance and officer
evaluations. Approximately 22 percent of the motorists tested received warnings
and five percent were cited for a motor vehicle violation other than DWI.

T ABLE 4
ESTIMATED AND M EASURED BAC (%) BY DISPOSITION, A GE CATEGORY, AND GENDER

Disposition & Estimated Measured

Category Number Percent BAC (Mean) BAC (Mean)

Warnings 65 21.9 0.060 0.044

Adults 57 0.063 0.045
Male Adults 53 0.063 0.044
Female Adults 4 0.070 0.054
Youth 8 0.036 0.038
Male Youth 6 0.037 0.038
Female Youth 2 0.035 0.040
Citations 15 5.1 0.055 0.046
Adults 11 0.050 0.040
Male Adults 9 0.047 0.043
Female Adults 2 0.065 0.029
Youth 4 0.070 0.062
Male Youth 2 0.060 0.055
Female Youth 2 0.080 0.070
Arrests 217 73.0 0.138 0.150
Adults 205 0.139 0.152
Male Adults 180 0.139 0.150
Female Adults 25 0.139 0.160
Youth 12 0.119 0.135
Male Youth 11 0.121 0.134
Female Youth 1 0.100 0.140
Total 297 100.0 0.117 0.122

-- 16 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

The data presented in Table 4 also show that officers tended to slightly over-
estimate the BACs of motorists who had lower BACs, and slightly under-estimate
BACs at the higher levels. Overall, however, officers’ estimates were extremely
accurate. Based on SFST results and officers’ observations, the officers’ mean
estimated BAC of the 297 motorists was 0.117 percent, compared to the mean
measured BAC of 0.122. Although statistically significant, the difference of 0.005
percent BAC is a trivial and operationally irrelevant under-estimate of actual BACs
that is within the margin of error of sophisticated evidentiary testing equipment.

CORRELATION A NALYSES
The accuracy of the SFSTs was further evaluated by conducting a series of
correlation analyses to identify the degree to which officers’ individual estimates of
BAC corresponded with subjects’ actual, or measured, BAC. A correlation coefficient
is a statistic, usually represented as r, that expresses the relatedness of two variables,
that is, the degree to which the variables co-vary. In this case, the two variables were
an officer’s estimate and the subject’s actual BAC. The Pearson product-moment
correlation method was used to calculate the relationship between these variables;
cases with complete SFST results (n=261) were used in this analysis.

If officers had predicted the precise BACs of all subjects (to three decimal
points), the correlation coefficient would be +1.00; the correlation coefficient would
be zero if there were no relationship between the estimated and actual BACs. For
predictive measures, especially those administered under field conditions, a
correlation of 0.65 to 0.70 is considered to be very high.

Table 5 presents the results of the correlation analyses. The table shows that
HGN test results had the highest correlation with measured BAC of the three
components of the SFST battery (r=0.65). However, a slightly higher correlation was
obtained when the results of the three component tests were combined (r=0.69). The
table also shows strong correlations between test results and officers’ estimated
BACs, indicating that officers were following procedures and interpreting test results
correctly. All of the correlations were found to be statistically significant (p=.005).

T ABLE 5
CORRELATIONS OF SFST S CORES TO ESTIMATED AND M EASURED BAC (%)
N=261 CASES WITH COMPLETE SFST S CORES

Correlation (r) Correlation (r)

with Estimated with Measured
Rank SFST(s) BAC BAC

1 3 Tests Combined 0.75 0.69

2 HGN 0.71 0.65
3 Walk-and-Turn 0.64 0.61
4 One Leg Stand 0.61 0.45

-- 17 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

DECISION A NALYSES
The third method used to evaluate the accuracy of the SFST battery was to
construct a decision matrix that describes the four possible combinations of the two
variables of interest, estimated and actual BACs above and below the levels of
interest. Figure 4 presents the first decision matrix, with the four major cells of the
matrix representing the four possible decisions at 0.08 percent BAC. The numbers in
the major cells are the number of cases for each type of decision out of the 297 SFST
administrations. The two shaded cells represent correct decisions based on SFST
results: 1) 210 motorists who officers estimated to have BACs equal to or greater
than 0.08 percent, who later were found to have BACs ≥0.08 by BAC testing (by
breath, blood, or urine analysis); and, 2) 59 motorists who officers estimated to have
BACs below 0.08 percent, who later tested below 0.08.
Figure 4 also reveals the incorrect decisions: 1) 24 motorists who officers
estimated to have BACs greater than 0.08 who later were found to have BACs below
that level (false positives); and, 2) four subjects who officers estimated to have BACs
below 0.08 who later tested above 0.08 (false negatives).
It can be calculated from the data contained in Figure 4 that officers’ decisions
were accurate in 91 percent of the 297 cases (i.e., [210+59]÷297=.906). Further, officers’
decisions to arrest were correct in 90 percent of the cases in which BAC was
estimated to be ≥0.08 (i.e., 210÷234=.897), and decisions not to arrest were correct in
94 percent of the cases in which BAC was estimated to be below 0.08 (i.e., 59÷63=.937).
These results indicate a high degree of accuracy, but it will be instructive to consider
more closely those cases in which incorrect decisions were made.

Officers' Estimated BACs

<0.08% ≥0.08%
≥0.08%

n=4 n=210 n=214

Measured BACs
<0.08%

n=59 n=24 n=83

n=63 n=234 N=297

Accurate in 91% of cases overall
90% accurate in "yes" decisions
94% accurate in "no" decisions

Figure 4. Decision matrix at 0.08 percent BAC.

-- 18 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

Table 6 presents a summary of the data for each of the 24 false positives (FPs).
These cases are labeled False Positives because the officers estimated the subjects’
BACs to be ≥0.08 percent, but subsequent testing found BACs below 0.08. However,
in several cases, officers were correct in identifying impairment, which probably
influenced their estimates of BAC.

T ABLE 6
S UMMARY OF FALSE POSITIVES

Number Is Estimate
Case Estimated of HGN Measured Consistent
Number BAC (%) Clues BAC (%) with Clues?

1 30 0.08 4 0.050 yes

2 34 0.08 4 0.058 yes
3 121 0.08 6 0.060 yes
4 186 0.08 4 0.063 yes
5 226 0.08 6 0.058 yes
6 227 0.08 4 0.060 yes
7 129 0.09 4 0.070 yes
8 175 0.09 4 0.070 yes
9 32 0.09 6 0.076 yes
10 127 0.09 6 0.028 yes
11 224 0.10 4 0.070 yes
12 16 0.10 6 0.070 yes
13 196 0.10 6 0.074 yes
14 52 0.11 4 0.050 yes
15 178 0.12 6 0.070 yes
16 246 0.12 6 0.069 yes
17 12 0.08 2 0.060 no
18 164 0.08 2 0.070 no
19 165 0.08 2 0.020 no
20 135 0.08 3 0.078 no
21 137 0.09 n/a 0.030 ?
22 75 0.09 2 0.048 no
23 104 0.09 3 0.037 no
24 13 0.12 0 0.043 no

In 16 of the cases listed in Table 6, the officers’ estimates of BAC were

consistent with the number of HGN clues observed (i.e., four or more HGN clues to

-- 19 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

support an estimate ≥0.08), however, the motorists subsequently were found to have
actual BACs below 0.08 percent. In seven of the cases, the officers’ estimated BACs
were inconsistent with the number of HGN clues observed. It is important to note
that six of the 24 false positives had measured BACs of 0.07 percent, and three had
BACs greater than 0.07 but less than 0.08 (i.e., 0.074, 0.076, and 0.078). All nine of
these BACs are within the margin of error of the testing devices. Further, Case
Number 16 was a juvenile (0.069), which rendered the difference between estimated
and measured BACs irrelevant in a zero tolerance jurisdiction; that is, it was a
correct arrest decision despite the BAC estimate. In addition, two of the subjects with
measured BACs of 0.07 were arrested for DWI, because the officers’ believed that
they were too impaired to be permitted to drive. Finally, Case Number 30, with an
estimated BAC of 0.08 and a measured BAC of 0.05 percent, was found to be a
psychiatric patient, which helped to explain her erratic behavior, poor SFST
performance, and apparent impairment.
Although the proportions of correct decisions presented in Figure 4 reflect a
high degree of accuracy, the accuracy of officers’ decisions is even better if some of
the borderline cases are accepted. An accuracy rate of 94 percent for all officer
decisions based on SFST results was calculated by including as correct decisions Case
16 (the youth with a 0.069 percent BAC) and the nine false positives with BACs
between 0.07 and 0.08, discussed in the previous paragraph.
Table 7 summarizes the four cases in which officers estimated the subjects’
BACs to be below 0.08 percent, but later found the measured BACs to be ≥0.08. Six
HGN clues would be expected for Case Number 193 (0.10 percent) and Case Number
99 (0.12 percent). It is unknown why the officers observed only two HGN clues. In
contrast, officers recorded four HGN clues for Case Number 131 and Case Number
114, which would indicate BACs greater than 0.08, however, the officers’ estimated-
BACs were only 0.06 percent. It is unknown why the officers did not follow the test
interpretation guidelines in these two cases; their low estimates probably reflect
other observations made in combination with SFST performance.

T ABLE 7
S UMMARY OF FALSE N EGATIVES

Number Is Estimate
Case Estimated of HGN Measured Consistent
Number BAC (%) Clues BAC (%) with Clues?

1 193 0.06 2 0.100 yes

2 99 0.06 2 0.120 yes
3 131 0.06 4 0.080 no
4 114 0.06 4 0.116 no

Similarly, in seven of the false positive cases listed previously in Table 6,

officers apparently did not follow the test interpretation guidelines; that is, fewer
than four HGN clues were reported, yet the officers’ estimated-BACs were at least

-- 20 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

0.08 percent. It is possible that other factors influenced the officers’ estimates. For
example, the subjects might have appeared to be more impaired than indicated by
HGN results as a consequence of prescription or recreational drugs taken in addition
to alcohol.
A series of decision analyses was performed to calculate the contributions of
the component tests of the battery to officers’ estimates of BAC. Figure 5 presents
three decision matrices, one for each of the SFSTs. The matrices are similar to the
one in Figure 4, but with the criterion numbers of clues at 0.08 percent BAC
substituted for officers’ estimates. Figure 5 shows the HGN test to be the most
accurate independent predictor of whether a motorist’s BAC is above or below 0.08
percent.

Number of HGN Clues

<4 ≥4
≥0.08%

n=4 n=205 n=209

Measured BACs
<0.08%

n=51 n=30 n=81

n=55 n=235 N=290

Accurate in 88% of cases overall
87% accurate in "yes" decisions
93% accurate in "no" decisions
Number of WAT Clues Number of OLS Clues
<2 ≥2 <2 ≥2
≥0.08%

≥0.08%

n=16 n=179 n=195 n=16 n=182 n=198

Measured BACs

Measured BACs
<0.08%

<0.08%

n=36 n=40 n=76 n=44 n=31 n=75

n=52 n=219 N=271 n=60 n=213 N=273

Accurate in 79% of cases overall Accurate in 83% of cases overall
82% accurate in "yes" decisions 86% accurate in "yes" decisions
69% accurate in "no" decisions 73% accurate in "no" decisions

Figure 5. Decision matrices at 0.08 percent BAC for each component test of the SFST battery.

-- 21 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

Further analyses were performed to explore methods for combining the

results of the three component tests. Only the 261 cases that included test results for
all three component tests could be used in this analysis. Of those cases, 73 were
found to have BACs below 0.08 percent and 188 cases had measured BACs ≥0.08
percent. In 162 of the 188 cases (86 percent), all three component SFSTs were
unanimous in their predictions.

Figure 6 presents a Venn diagram that illustrates the contributions of the

three tests to the 14 percent of cases in which a discrepancy occurred. The figure
shows there were 162 cases with BACs ≥0.08 in which all three SFSTs indicated a
BAC ≥0.08 (the number outside the circles in Figure 6), and 26 cases in which one or
more test disagreed (the numbers inside the circles). A single test indicated a BAC
below 0.08 in 17 of the cases (8+2+7), and two tests were involved in nine of the
cases (1+1+7). There were no cases in which all three tests predicted incorrectly.

HGN
2

1 1
0
8 7 7
WAT OLS

162 Cases ≥0.08% BAC in which all three tests agree

(represented by the area outside the three circles)

Figure 6. Venn diagram of 188 cases ≥0.08% BAC; 26 cases in which all three tests do not agree.

The horizontal gaze nystagmus test (HGN in the diagram) was about four
times less likely to be the source of a discrepancy than the other two tests. Only two
of the single-test discrepancies were attributable to HGN results, compared to eight
cases for the Walk and Turn test (WAT), and seven cases for the One Leg Stand
(OLS). Overall, the HGN test was involved in only four of the discrepancies,
compared to 16 cases for the Walk and Turn and 15 cases for the One Leg Stand.

The question of the SFST battery’s accuracy in discriminating BACs above and
below 0.04 percent is addressed by the following decision matrix, presented in Figure
7; the shaded cells of the matrix again represent correct decisions based on SFST
results. The figure shows that officers estimated motorists’ BACs to be equal to or
greater than 0.04 but under 0.08 percent in 54 cases, and in 51 of those cases their
estimates were found to be correct by subsequent breath, blood, or urine testing;

-- 22 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

these values result in an accuracy rate of 94 percent for these decisions (i.e.,
51÷54=.94). The figure also shows that officers estimated that 29 motorists had BACs
below 0.04, and in 15 of those cases their estimates were found to be correct by
subsequent testing, resulting in a 52 percent accuracy rate (15÷29=.52). Overall,
officers were accurate in 80 percent of the cases when discriminating between
subjects who were above 0.04 but below 0.08 percent BAC (i.e., [51+15]÷83=.80).

Officers' Estimated BACs

≥0.04 <0.08% <0.04% ≥0.04 <0.08%

n=14 n=51 n=65

Measured BACs
<0.04%

n=15 n=3 n=18

n=29 n=54 N=83

Accurate in 80% of cases overall
94% accurate in "yes" decisions
52% accurate in "no" decisions

Figure 7. Decision matrix at 0.04 percent BAC.

E VALUATION OF SFST ACCEPTABILITY

In interviews and during ride-along observations, the officers who
participated in the study fully accepted the SFSTs for evaluating motorists for DWI
at BACs below 0.10 percent. All of the officers were formally trained in SFST
administration and scoring and all had sufficient field experience to develop
confidence in their abilities to discriminate at the 0.08 level. Further, it was the
officers’ experience with the SFST battery that the component tests could be
administered to all but a small proportion of drivers and under all reasonable
environmental conditions.

Interviews also were conducted with representatives of the San Diego City
Attorney’s Office to inquire concerning the acceptability of the SFSTs to prosecutors
and judges in DWI cases. The attorneys interviewed reported that none of the 298
DWI arrests made by participating officers during the study period was negatively
affected by the SFST battery, or by excluding the other tests that traditionally had
been used by the department.

-- 23 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

The attorneys further explained that as prosecutors they normally prefer as

much evidence as possible, and in a DWI case more tests usually generate more
evidence they can use. However, it has been their recent experience that a test used
by another local law enforcement agency has negatively affected cases they have
prosecuted. Defense attorneys have been unsuccessful in their challenges of
NHTSA’s SFST battery, but they have successfully challenged the validity of the
other test because it has not been evaluated in a systematic and scientific manner.
Prosecutors who were interviewed suggested that the optimum situation would be
for all law enforcement agencies to restrict their field sobriety evaluations to the
same standardized battery of three tests.

-- 24 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

IMPLICATIONS

The research documented in this report found that NHTSA’s Standardized

Field Sobriety Test Battery accurately and reliably assists officers in making DWI
arrest decisions at 0.08 percent BAC. The study also found that the SFSTs can be used
to assist officers in making arrest decisions at 0.04 percent BAC by using two HGN
clues as the criterion rather than four clues, which is the criterion for a 0.08 percent
or above BAC determination. The primary implication of the study results is that
the SFST battery is a valid method for making roadside DWI decisions at 0.08 and
0.04 percent BAC. Specific implications of the study results are presented in the
following paragraphs in response to the research questions listed previously.

HOW ACCURATELY DO THE TESTS DISCRIMINATE BETWEEN S UBJECTS WHO

ARE ABOVE OR BELOW 0.08 AND 0.04 PERCENT BACS?
This study found NHTSA’s SFST battery to be an accurate method for
discriminating motorists’ BACs above and below 0.08 percent and above and below
0.04 percent, when the tests are conducted by trained officers, as summarized below.

COMPARISON OF MEANS
The mean estimated BAC of the 297 motorists included in the study was 0.117
percent, compared to the mean measured BAC of 0.122. The difference of 0.005
percent BAC (i.e., five one-thousandths of a percent BAC) is very small and
operationally irrelevant. The accuracy of officers’ estimates during this study, in
large measure, confirms the anecdotal accounts and observations of officers in the
field that suggest remarkable abilities to predict a motorists’ BAC on the basis of
SFST results.

CORRELATION A NALYSES
Correlation analyses found the HGN test to be very predictive of measured
BACs (r=0.65). A higher correlation was obtained when the results of the three
component tests were combined (r=0.69). All of the correlations are statistically
significant, meaningful, and in the rank order expected from previous SFST
research.

DECISION A NALYSES
Decision analyses found that officers’ estimates of whether a motorist’s BAC
was above or below 0.08 or 0.04 percent were extremely accurate. Estimates at or
above the 0.08 level were accurate in 91 percent of the cases, or as high as 94 percent
if explanations for ten of the false positives are accepted. Estimates at or above the
0.04 level (but below 0.08) were accurate in 94 percent of the relevant cases. It is
important to note that officers’ decisions not to arrest were more accurate at 0.08
than at 0.04 (94 percent compared to 52 percent).

Although the relatively small number of low BACs in the data base (n=83)
might constrain confidence in the SFSTs at the 0.04 level, the data strongly suggest

-- 25 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

operational utility to accurately discriminate above or below 0.04 percent BAC.

Further, these results are consistent with the results of a recent study conducted to
evaluate the SFST battery for use by officers in Colorado.

Colorado has a two-tier statute that permits officers to arrest motorists for
driving under the influence (DUI) if found to have a BAC ≥ 0.10 percent, and for a
lesser offense, driving while ability impaired (DWAI), if found to have a BAC ≥0.05
but below 0.099 percent. Of the 234 drivers tested during the Colorado study for
whom BACs were known, 93 percent of the officers’ decisions to arrest at the 0.05
percent criterion were correct, and 64 percent of the decisions to release were correct.
Overall in the Colorado study, 86 percent of the officers’ decisions at the 0.05 level
were correct, based on SFST results (Burns and Anderson, 1995; Anderson and
Burns, 1997).

WHICH OF THE C OMPONENTS OF THE SFST B ATTERY IS/ARE THE BEST

PREDICTOR(S) OF BAC?
The horizontal gaze nystagmus test was found to be the most predictive of the
three component tests, but correlations with measured BACs were higher when the
results of all three tests were combined, as reported earlier. The implications of this
study result are that all components of the SFST battery should be administered
when possible or practical. However, the data indicate that the HGN test alone can
provide valid indications to support officers’ arrest decisions at both 0.08 and 0.04
percent BAC.

HOW RELIABLE , OR C ONSISTENT, ARE THE TESTS ?

Reliability is a measurement concept that represents the consistency with
which a test measures a type of performance or behavior. In the current context, a
reliable field sobriety test provides consistent results when administered to the same
individual by two different officers, under nearly identical conditions. This type of
“inter-rater” reliability was impossible to measure directly during this study, due to
the constraints imposed by field conditions. In particular, it would have been
unrealistic to subject motorists to the SFST battery twice, or to require that officers
operate in pairs during their patrols.

Evidence of SFST reliability can be found in the results of the previous

laboratory studies, in which the constraints on repeated measure were eliminated by
the use of paid subjects and officers. Tharp, Burns, and Moskowitz (1981) found
relatively high inter-rater reliability for BAC estimates based on SFST results (r=.72).
The researchers also found that inter-rater reliability increased in subsequent
sessions (r=.80), indicating the important role of training and experience in
achieving accuracy, reliability, and overall proficiency.

In addition, correlation coefficients, in general, are measures of reliability. For

this reason, the correlations between estimated and actual BACs obtained during the
field study (r=.69) indicate a high degree of reliability for tests designed to be
administered at roadside.

-- 26 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

ARE THE TESTS USABLE BY POLICE OFFICERS UNDER A VARIETY OF ROADSIDE

C ONDITIONS? ARE THEY READILY ACCEPTED BY OFFICERS AND PROSECUTORS ?
All of the officers who participated in this study were members of the San
Diego Police Department’s alcohol enforcement unit, all had previously received
NHTSA-approved training in DWI detection and SFST administration, and all had
at least three years of experience in the Traffic Division before joining the special
unit. Prior to beginning the field study, the officers demonstrated competence in the
administration of the component tests and interpretation of test results.
Participation was limited to members of the alcohol-enforcement unit of a single
law enforcement agency. These experience and training requirements were
imposed, to control variables, to the extent possible, that might affect study results.

As a consequence of the selection criteria, all participating officers were

proficient in the use of the SFST battery. The officers reported that they use their
SFST skills daily in their work, and their experience has made them confident in the
ability of the test battery to discriminate at 0.08 percent BAC, and at lower levels.
Further, officers reported that the tests can be administered in all reasonable
environmental conditions. In short, the officers who participated in this study
consider the SFST battery to be extremely useful, in fact, essential tools for the
performance of their professional duties.

The prosecutors interviewed during the study reported that the SFST battery
has been acceptable to them and the courts because it was developed and validated
in a systematic and scientific manner. They suggested that all law enforcement
agencies should limit officers to use of the SFST battery in performance evaluations
of DWI because other tests usually lack credibility in court. No problems were
experienced in any of the 298 cases resulting from the field study, indicating the
SFSTs to be fully acceptable to the courts in establishing probable cause to arrest a
motorist for DWI.

N OTE A BOUT THE A CCEPTABILITY OF THE HGN TEST

Many law enforcement officers from across the United States have reported
their sincere appreciation to NHTSA for developing the SFST battery, and in
particular, the horizontal gaze nystagmus test. However, some officers have
expressed frustration about the resistance of some courts to accept HGN results,
despite the clear and unequivocal support of scientific research and field experience.
It is likely that this remaining resistance to the horizontal gaze nystagmus test is
attributable to a misunderstanding concerning the purpose of a field sobriety test,
and can be explained by reference to “face validity,” a term used in the behavioral
sciences to describe one component of a measure’s acceptability.

Many individuals, including some judges, believe that the purpose of a field
sobriety test is to measure driving impairment. For this reason, they tend to expect
tests to possess “face validity,” that is, tests that appear to be related to actual driving
tasks. Tests of physical and cognitive abilities, such as balance, reaction time, and
information processing, have face validity, to varying degrees, based on the

-- 27 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

involvement of these abilities in driving tasks; that is, the tests seem to be relevant
“on the face of it.” Horizontal gaze nystagmus lacks face validity because it does not
appear to be linked to the requirements of driving a motor vehicle. The reasoning is
correct, but it is based on the incorrect assumption that field sobriety tests are
designed to measure driving impairment.

Driving a motor vehicle is a very complex activity that involves a wide

variety of tasks and operator capabilities. It is unlikely that complex human
performance, such as that required to safely drive an automobile, can be measured at
roadside. The constraints imposed by roadside testing conditions were recognized by
the developers of NHTSA’s SFST battery. As a consequence, they pursued the
development of tests that would provide statistically valid and reliable indications
of a driver’s BAC, rather than indications of driving impairment. The link between
BAC and driving impairment is a separate issue, involving entirely different
research methods. Those methods have found driving to be impaired at BACs as
low as 0.02 percent, with a sharp increase in impairment at about 0.07 percent
(Moskowitz and Robinson, 1988; Stuster, 1997). Thus, SFST results help officers to
make accurate DWI arrest decisions even though SFSTs do not directly measure
driving impairment.

Horizontal gaze nystagmus is the most accurate diagnostic of BAC available to

officers in the field. HGN’s apparent lack of face validity to driving tasks is
irrelevant because the objective of the test is to discriminate between drivers above
and below the statutory BAC limit, not to measure driving impairment.
Throughout the United States, DWI laws permit arrest decisions to be made on the
basis of the statutory BAC limit, irrespective of a specific motorist’s degree of
impairment. Motorists also can be arrested at BACs below the statutory limit if their
driving performance is demonstrably impaired by alcohol or other drugs.

C ONCLUSIONS
The results of this study provide clear evidence of the validity of the
Standardized Field Sobriety Test Battery to discriminate above or below 0.08 percent
BAC. Further, study results strongly suggest that the SFSTs also accurately
discriminate above or below 0.04 percent BAC.

Finally, in addition to establishing the validity of the SFST battery, this study
has found the tests to be acceptable, indeed welcomed, by law enforcement officers
and DWI prosecutors.

-- 28 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

REFERENCES

Anderson, E.W. and Burns, M. (1997). Standardized Field Sobriety Tests: A Field
Study. Proceedings of the 14th International Conference on Alcohol, Drugs
and Traffic Safety Volume 2, 635-639.

Aschan, G. (1958). Different types of alcohol nystagmus. Acta Otolaryngology,

Supplement 140, 69-78.

Burns, M. and Anderson, E.W. (1995). A Colorado Validation Study of the

Standardized Field Sobriety Test (SFST) Battery. Colorado Department of
Transportation.

Burns, M. and Moskowitz, H. (1977). Psychophysical Tests for DWI Arrest. U.S.
Department of Transportation, National Highway Traffic Safety
Administration Final Report DOT-HS-5-01242, Washington, D.C.

Harris, D.H., Dick, R.A., Casey, S.M.,and Jarosz, C.J. (1980). The Visual Detection of
Driving W h i l e Intoxicated. U.S. Department of Transportation, National
Highway Traffic Safety Administration Final Report DOT-HS-7-1538.

Harris, D.H. (1980). Visual detection of driving while intoxicated. Human Factors,
22(6), 725-732.

Lehti, H.M.J. (1976). The effects of blood alcohol concentration on the onset of gaze
nystagmus. Blutalkohol, Vol. 13, 411-414.

Moskowitz, H., and Robinson, C.D. (1988). Effects of Low Doses of Alcohol o n
Driving-Related Skills: A Review of the Evidence. U.S. Department of
Transportation, National Highway Traffic Safety Administration, DOT-HS-
807-280, Washington, D.C.

Pentilla, A., Tenhu, M., and Kataja, M. (1971). Clinical Examination For Intoxication
In Cases of Suspected Drunken Driving. Statistical and Research Bureau of
TALJA. Iso Roobertinkatu 20, Helsinki, Finland.

Stuster, J.W. (1997). The Detection of DWI at BACs Below 0.10. U.S. Department of
Transportation, National Highway Traffic Safety Administration Final Report
(in press).

Stuster, J.W. (1993). The Detection of DWI Motorcyclists. U.S. Department of

Transportation, National Highway Traffic Safety Administration Final Report
DOT-HS-807-839.

-- 29 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

Tharp, V., Burns, M., and Moskowitz, H. (1981). Development and Field Test of
Psychophysical Tests for DWI Arrest. U.S. Department of Transportation,
National Highway Traffic Safety Administration Final Report DOT-HS-805-
864, Washington, D.C.

Wilkinson, I.M.S., Kime, R., and Purnell, M. (1974). Alcohol and human eye
movement. Brain, 97, 785-792.

-- 30 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

APPENDIX A

S TANDARDIZED F IELD S OBRIETY TESTING

-- 31 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

S TANDARDIZED F IELD S OBRIETY TESTING

The Standardized Field Sobriety Test (SFST) is a battery of three tests adminis-
tered and evaluated in a standardized manner to obtain validated indicators of
impairment and establish probable cause for arrest. These tests were developed as a
result of research sponsored by the National Highway Traffic Safety Administration
(NHTSA) and conducted by the Southern California Research Institute. A formal
program of training was developed and is available through NHTSA to help police
officers become more skillful at detecting DWI suspects, describing the behavior of
these suspects, and presenting effective testimony in court. Formal administration
and accreditation of the program is provided through the International Association
of Chiefs of Police (IACP). The three tests of the SFST are:
• Horizontal gaze nystagmus (HGN),
• Walk-and-turn, and
• One-leg stand.
These tests are administered systematically and are evaluated according to
measured responses of the suspect.

HGN TESTING
Horizontal gaze nystagmus is an involuntary jerking of the eye which occurs
naturally as the eyes gaze to the side. Under normal circumstances, nystagmus
occurs when the eyes are rotated at high peripheral angles. However, when a person
is impaired by alcohol, nystagmus is exaggerated and may occur at lesser angles. An
alcohol-impaired person will also often have difficulty smoothly tracking a moving
object. In the HGN test, the officer observes the eyes of a suspect as the suspect
follows a slowly moving object such as a pen or small flashlight, horizontally with
his or her eyes. The examiner looks for three indicators of impairment in each eye:
if the eye cannot follow a moving object smoothly, if jerking is distinct when the eye
is at maximum deviation, and if the angle of onset of jerking is within 45 degrees of
center. If, between the two eyes, four or more clues appear, the suspect likely has a
BAC of 0.10 or greater. NHTSA research indicates that this test allows proper
classification of approximately 77 percent of suspects. HGN may also indicate
consumption of seizure medications, phencyclidine, a variety of inhalants,
barbiturates, and other depressants.

WALK AND TURN

The walk-and-turn test and one-leg stand test are “divided attention” tests
that are easily performed by most unimpaired people. They require a suspect to
listen to and follow instructions while performing simple physical movements.
Impaired persons have difficulty with tasks requiring their attention to be divided
between simple mental and physical exercises.
In the walk-and-turn test, the subject is directed to take nine steps, heel-to-toe,
along a straight line. After taking the steps, the suspect must turn on one foot and

-- 32 --
Final Report
Validation of the SFST Battery at BACs Below 0.10 Percent

return in the same manner in the opposite direction. The examiner looks for eight
indicators of impairment: if the suspect cannot keep balance while listening to the
instructions, begins before the instructions are finished, stops while walking to
regain balance, does not touch heel-to-toe, steps off the line, uses arms to balance,
makes an improper turn, or takes an incorrect number of steps. NHTSA research
indicates that 68 percent of individuals who exhibit two or more indicators in the
performance of the test will have a BAC of 0.10 or greater.

ONE LEG S TAND

In the one-leg stand test, the suspect is instructed to stand with one foot
approximately six inches off the ground and count aloud by thousands (One
thousand-one, one thousand-two, etc.) until told to put the foot down. The officer
times the subject for 30 seconds. The officer looks for four indicators of impairment,
including swaying while balancing, using arms to balance, hopping to maintain
balance, and putting the foot down. NHTSA research indicates that 65 percent of
individuals who exhibit two or more such indicators in the performance of the test
will have a BAC of 0.10 of greater.

C OMBINED M EASURES
NHTSA’s SFST training materials instruct officers in the use of the following
decision table for combining the results of the HGN and Walk and Turn test.
HGN Clues
0 1 2 3 4 5 6
0

1
2
Walk and Turn Clues

3
4

9
Along the top of the table, circle the number of the
subject's HGN clues. Along the left side of the table,
circle the number of the subject's Walk and Turn clues.
Draw a line down from the number of HGN clues and
a line across from the number of Walk and Turn clues.
If the intersection is within the shaded area, the subject
has a BAC ≥0.10 percent.

-- 33 --
OVERHEAD(S)
HANDOUT(S)
 SFST ONE-DAY REFRESHER COURSE
PARTICIPANT MANUAL
COLLATING PRINT GUIDE

1. Agenda
Agenda.SFST One-day Refresher.pdf

2. Preface
Preface.pdf

3. Introduction and Overview

Introduction and Overview.pdf

4. Session One: Phase One: Vehicle in Motion & Phase Two: Personal Contact
Session 1.Part 1.Phase 1.pdf
Session 1.Part 2.Phase 2.pdf

5. Session Two: Concepts and Principles of the Standardized Field Sobriety

Tests & Phase Three: Pre-Arrest Screening
Session 2.Part 1.Concepts & Principles.pdf
Session 2.Part 2.Phase 3.pdf

6. Session Three: Test Battery Demonstrations & Dry Run Practice Session
Session 3.part 1.Demo.pdf
Session 3.Part 2.Dry Run.pdf

7. Sessions 4 & 5: “Testing Subjects” Practice – Dry Lab – Video subjects

Session 4 & 5.Dry Lab.pdf

8. Attachments
ATTACHMENT COVER PAGE.pdf
Attach 1.Detection Below.10.pdf
Attach 2.Colorado Validation Study.pdf
Attach 3.Florida Validation Study.pdf
Attach 4.Detection Below .10.pdf

9. Overheads (optional based on instructor preferences)

Into & Overview.ppt
Session 1.Part 1.Phase 1.ppt
Session 1.Part 2.Phase 2.ppt
Session 2.Part1.Concepts & Principles.ppt
Session 2.Part 2.PreArrestScreen.ppt
Session 3.Demo.ppt
Session 4& 5.Dry Lab.ppt
10. Handouts (Specific to each instructor)
HANDOUT COVER PAGE.pdf