If you have issues viewing or accessing this file contact us at NCJRS.gov.

THE VI" I ' ~ A L DETECTION OF DRIVING WHILE INTOXICATED

Field Test of Visual Cues and Detection Methods

Douglas H. Harris Robert A. Dick Steven M. Casey Christopher J. Jarosz ANACAPA SCIENCES, INC. P.O. Drawer Q Santa Barbara, California 93102

Contract DOT-HS-7-1538:$271,834

,WILL
April 1980 FINAL REPORT Document is available to the U.S. public through the National Techn ical Information Service Springfield, Virginia 22161 . prepared for

_/

",

,,'~h

U.S. DEPARTMENT OF TRANSPORTATION..

N A T I O N A L HIGHWAY T R A F F I C S A F E T Y A D M I N I S T R A T I O N

Office of Driver and Pedestrian Research.

Washington, D.C. 20590
/'

"~

"

It

NOTICE This document is disseminated under the sponsorship of the Department of Transportation in the interest of information exchange. The United States Government assumes no liability for the contents or use thereof.
q

".:-!

p~b

"

":

,:I..

t.-,,

"

AC,QU StIT ON

~S
~i~ ~i ,~ i i ~~.ii~. ~.~

TECHNICAL SUMMARY
CONTRACTOR Anacapa Sciences, Inc. REPORT Tri'hE CONTRACT NUMBER DOT-HS-7-1538

REPORT DATE THE VISUAL DETECTION OF DRIVING WHILE April 1980 INTOXICATED Field Test of Visual Cues and Detection Methods
-

REPORT AUTHOR(S) Douglas H. Harris, Robert A. Dick, Steven M. Casey, Christopher J. Jarosz Onthe-road detection of driving while intoxicated (DWI) involves the observation and interpretation of visual cues by police patrol officers. The effectiveness of DWI detection depends not only on the frequency with which patrol officers see and recognize cues indicative of DWI, but on the extent to which observed cues discriminate between DWI and driving while sober (DWS). This research project was conducted to answer the following questions: What visual cues occur frequently enough to be useful for DWI detection? To what extent do different cues discriminate between DWI and DWS? How can information on cue occurrence and discriminability be used best for on-theroad detection of DWI? In the first phase of the project, reported earlier, the literature was reviewed, DWI detection experts were interviewed, a large sample of arrest reports was analyzed, and an on-the-road study of DWI detection was conducted to obtain data on the relative discriminability and frequency of occurrence of visual detection cues. The end product of the first phase was a set of conclusions about DWI detection, and a prototype DWI detection guide designed to facilitate application of the research findings to on-the-road detection of DWI. In the second phase of the project, reported here, a DWI Detection Guide and an explanatory booklet were developed and tested. The Guide was a small card of white plastic printed with blue. The field test of the Guide was conducted with a sample of i0 law enforcement agencies at locations throughout the United States. The test was designed to provide both longitudinal and cross-secti0nal analyses of several measures likely to reflect the impact and utility of using the Guide, and to verify the values contained in the Guide.

(Continued) PREPARED FOR THE DEPARTMENT OF TRANSPORTATION, NATIONAL HIGHWAY TRAFFIC SAFETY ADMINSTRATION, UNDER CONTRACT NO. DOT-HS-7-1538. THE OPINIONS, FINDINGS, AND CONCLUSIONS EXPRESSED IN-THIS PUBLICATION ARE THOSE OF THE AUTHORS AND NOT NECESSARILY THOSE OF THE NATIONAL HIGI-BVAY TRAFFIC SAFETY ADMINSTRATION.

Use of the Guide was accompanied by a statistically significant overall increase in DWI arrest rate of 12 percent Rates during a 3-month period, in which the Guide was used, were compared to a 12-month baseline period. Five individual agencies had significant increases of up to 94 percent; four agencies had no change; and one agency had a significant decrease. Although there were no statistically significant changes in detection practices reflected by greater use of the more discriminating cues or by arrests of drivers with lower BAC levels, trends were in those directions. Experienced police officers who used the Guide expressed doubts about its value in improving their own DWI enforcement ability. However, they considered the Guide to be valuable for increasing patrol sensitivity to important cues, training new patrol officers, preparing DWI arrest reports, and supporting court testimony Field-test results led to a couple of minor modifications in the Guide and the explanatory booklet DWI detection probabilities were based on combined data from the early detection study and from the field test, thus providing a data base of 4662 detection events. Several other minor modifications were made to further simplify and clarify the Guide The resulting DWI Detection Guide is shown below.

DWI DETECTION GUIDE

C h a n c e s in 100 of n i g h t t i m e driver with BAC e q u a l or g r e a l e r t h a n .10

TURNING WITH WIDE RADIUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . . . . . . APPEARING TO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ALMOST STRIKING OBJECT OR VEHICLE . . . . . . . . . . . . . . . . . . WEAVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DRIVING ON OTHER THAN DESIGNATED ROADWAY . . . . . . . SWERVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SLOW SPEED (MORE THAN 10 MPH BELOW LIMIT) . . . . . . STOPPING (WITHOUT CAUSE) IN TRAFFIC LANE . . . . . . . . FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DRIFTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TIRES ON CENTER OR LANE MARKER . . . . . . . . . . . . . . . . . . . . . . BRAKING ERRAT CALLY DRIVING INTO OPPOSING OR CROSSING TRAFFIC . . . . . . . . . . SLOW RESPONSE TO TRAFFIC SIGNALS . . . . . . . . . . . . . . . . . .

65 65 GO GO 60 55 55 50 50 50 50 45 45 45 40

SIGNALLING INCONSISTENT WITH DRIVING ACTIONS . . . . . 40 STOPPING INAPPROPRIATELY (OTHER THAN IN LANE) . . . ; 35 TURNING ABRUPTLY OR ILLEGALLY . . . . . . . . . . . . . . . . . . . . . . 35 ACCELERATING OR DECELERATING RAPIDLY . . . . . . . . . . . . . . HEADLIGHTS OFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Specia~ a d i u s t m e n t to the c u e v a l u e s 2 o r m o r e cues observed: a dd 10 to the larger value B A C e qua l to or gr e a te r than .05: add 15 to the v a l ue obtained f o r B A C equa] 11oo r greater t h a n .10

30 30

A short 16-mm sound film in color was produced to describe the detection cues and introduce the Guide

I. I~*part No.

Technical Report Documentation Page

i ~. ~ovefnment Accession NO. 3. Recipient~'s Catalog No.

4. T,Ile and Subtitle

5. Report Dale

THE VISUAL DETECTION OF DRIVING WHILE INTOXICATED ~ield Test of Visual Cues and Detection Methods 7. A~,hor's) Douglas H. Harris, Robert A. Dick, Steven M. Casey, and Christopher J. Jarosz
9, Pertarming Org~izet~on Name ~ d Address

;April

1980

6. Performing Organization Code

8. Performing Organization Report No.

32]-2
10. Work Unit No. (TRAIS) | 1. Contract or Grant No.

Anacapa Sciences, Inc. P.O. Drawer Q Santa Barbara, C a l i f o r n i a

12. Sponsoring Agency Name and Address

93102

DOT-HS-7-1 538
13. Type of Report and Period Covered

16. Abstract

Final Report U.S. Department of Transportation National Highway T r a f f i c Safety Administration February 1979 - May 1980 O f f i c e of Driver and Pedestrian Research ]4. Sponsoring Agency Code Washington, DC 20590 is. s~ppl . . . . ,a,~Not., Interim report: Harris, D. H., Howlett, J. B. and Ridgeway, R G. The visual detection of driving while intoxicated, project interim report: Identification of visual cues and development of detection methods. Anacapa Sciences, Inc. f o r NHTSA~ January 1979. (PB 80 108 327) A Drunk Driver Detection Guide was developed and tested at a sample of I0 law enforcement agencies at locations throughout the United States. The t e s t was designed to provide both longitudinal and cross-sectional analyses of several measures l i k e l y to r e f l e c t the impact and u t i l i t y of using the Guide, and to v e r i f y the values contained in the Guide. Use of the Guide was accompanied by a s t a t i s t i c a l l y s i g n i f i c a n t overall increase in DWI arrest rate of 12 percent. Rates during a three-month period, in which the Guide was used, were compared to a 12-month baseline period. Five individual agencies had s i g n i f i c a n t increases of up to 94 percent; four agencies had no change; and one agency had a s i g n i f i c a n t decrease. The DWI p r o b a b i l i t y values contained in the Guide were v e r i f i e d by the f i e l d - t e s t r e s u l t s . Average values obtained during the f i e l d t e s t were e s s e n t i a l l y the same as average values on the Guide, and Guide values f o r individual cues correlated s i g n i f i c a n t l y with corresponding values calculated from f i e l d - t e s t data. Experienced police o f f i c e r s who used the Guide expressed doubts about i t s value in improving t h e i r own DWI enforcement a b i l i t y . However, they considered the Guide to be valuable f o r increasing patrol s e n s i t i v i t y to important cues, t r a i n i n g new patrol o f f i c e r s , preparing DWI arrest reports, and supporting court testimony.

|7% Key Words

Detection Drinking Drivers DWI Enforcement Visual Cues

18. Distribution Statement

Document is available to the U.S. public through the National Technical Information Service, S p r i n g f i e l d , V i r g i n i a 22161
20. Security Classif. (of this page)

]9. Sec~rtty C|assl|. (ol this report)

-2|. No. of Pages

22. Price

UNCLASSI FI ED
Form DOT F 1700.7 (e-72)

UNCLASSIFIED
Reproduction of completed page authorized

ill/iv

PREFACE
This is the second of two reports prepared under Contract DOT-HS-701538 on the visual detection of d r i v i n g while intoxicated. report described the i d e n t i f i c a t i o n of visual detection methods that led to a Drunk Driver Detection Guide. The i n i t i a l This report cues and development of

describes the Guide, and the f i e l d test conducted to evaluate and v e r i f y the Guide. The study involved the p a r t i c i p a t i o n of 10 d i f f e r e n t police agencies; without the cooperation and support of these agencies the work would not have been possible. We are grateful f o r the exceptional contributions to the project of the administrative and patrol personnel of these agencies. The agencies, along with our p r i n c i p a l point of contact, are l i s t e d below in alphabetical order. Albuquerque (New Mexico) Police Department: Lieutenant John Nelson Englewood (Colorado) Police Department: Captain Allan stanley Eugene (Oregon) Police Department: Sergeant Robert Laws Captain James Kleeman Captain Robert Wilsey Lieutenant Jack Nelson Evansville (Indiana) Police Department:

Monroe County (New York) S h e r i f f ' s Office: Santa Ana ( C a l i f o r n i a ) Police Department: Tacoma (Washington) Police Department: Topeka (Kansas) Police Department:

Pulaski County (Arkansas) S h e r i f f ' s Office: Lieutenant Jim Kemmeling Captain Phil Sessions Sergeant Jim Fravel

Major Dan Mallory

Vanderburgh County (Indiana) S h e r i f f ' s Office:

The Contract Technical Manager of this phase of the project was Mr. William C. Wheeler, Jr.; we are appreciative of the assistance and support he provided. Data processing was designed and conducted by Mr. Curtiss Mosso, Computer Center, University of C a l i f o r n i a at Santa Barbara, using the S t a t i s t i c a l Analysis System (SAS).

r \

METRIC C0~VEilSION FACTORS

Approximate Conversions to Metric Measures Approximate
- -

Conversions

from by

Metric

Measures To Find Symbol

et Symbol
el

Symbol

W h x n You Know

Multiply

by

To Find

Symbol
. - m - -

W h x n You K n o w

Multiply

LENGTH

[ENGTtl
.

~
. _ m

nvn

in f! yd mi

inches feel yards miles

"2.5 30 0.9 1.6 AREA

Centimelers ee.limeter$ meters kilo~nelerS

cm cm m km

~--

km

millin~Jters centimeters mututs mt~terS kihxneters

0.04 0.4 3.3 1.1 0.6

inches i.ches feet yards miles

in in ft yd mi

"'2

-Z

z
~

~
AREA. cm 2 nl 2 km 2 i,
he

in2 tt 2 yd '~ n, 2 <:

sq*lare Sq;lare equals square acres

inches IIJul yald$ miles

6,5 0.09 0.8 2.6 0.4 MASS (weight)

sq.aro contimefore square motels Sq~are ,~.!l(,r s Sq.a,e k i l ~ . u t e r s hectares

6182 nl 2 m2
km 2

st,

" -

~
__ U3

square centimeters square n~lore square kil~nolors hectares (10,000 m 2)

O.IG 1.2 0.~ 2.S

square inches squale yazds squale miles acres

in x yd 2 ml 2

ha MASS
_ _ et

(weight) 0.035 2.2 1.1 Ounces po, md S. short to~s

ol Ib

ounces po..ds sh ot't Ions C2000 Ib}.

28 0.45 O.g VOtUME

grams kdog;ame , Ionnos

g kg I

grams kilog,ame tonnes (I000 kg)

VOLUME milliliiees m,llitlters millililels liters li(ers filets liters cubic n,etels cubic meturs (exact)
~-~" c

tsp Tbsp fl e l c pt qt . pal ft 3 Yd J

teaspoons tablespoons fluud minces cups pmlS IlU~HS gallo.s cubic t~et cubic yards

5 15 30 0,24 0.47 0.~5 3.8 0.03 O.7G

ml ml ml I I t I m3 m3

-.
c~

-_--~_ ~ --

=o

ml I

.|
i

....

::

m]

millililerl lltefs lilex s lilelS cllblc molars C|lflic (TI~IOIS

0.03 2.1 1.06 0.26 35 l .~

.: !' '

fluid ounces pints quails galloczs cubic feet CUI}ic y~lld$

fl os pt qt ~=gal if) vd z

.-..2-"-'2

~= ~

TEMPERATURE Celsius tu.lperafuro 9/5 Hhen add 32)

(exact) Fahrenheit temperature oF 2p2

TEMPEI1ATU[?E Fahrenheit temperature

5/9 (;~fler subtracting 32)

Celsius te,,~pefature

~C OF -40 0 ]._.,l~L~l ..... I;

--40 oC -?0

32 iI i
0 ~0

98.6

t ,n k 2.54 f,,.j, tl)l. l ,~ ,,ll.., i,.,t, t c,.,v,.l~,,....i..l .,~.,, it,,t,t,h,d t,lbu,~, ~e*, fJUS ;.t,~,l. P.bl..~Hej,

.~" u~

60 3r

,!

B10

IO0 C

TABLE OF CONTENTS
Page TECHNICAL SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . . -. . . . . . . . . . . 1 5 6 10 12 12 15 15
. 21 .

Background . . . . . . . . . . . . . . . . . . . . . . . . Field-Test Objectives . . . . . . . . . . . . . . . . . . . METHOD . . . . . . ~ . . . . . . . . . . . . . . . . . . . .

Field-Test Design . . . . . . . . . . . . . . . . . . . Participating Agencies . . . . . . . . . . . . . . . . . . . Field-Test Procedures . . . . . . . . . . . . . . . . . . . .

RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I m p a c t on DWI A r r e s t R a t e s . . . . . . . . . Changes i n DWI D e t e c t i o n P r a c t i c e s . . . . . Verification of the Guide . . . . . . . . . Opinions of Guide Utility. . . . . . . . . . CONCLUSIONS AND RECOMMENDATIONS
APPENDIX

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . .

21 24 26 34 39
A-I

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

LIST OF FIGURES
Figure I 2 3 FIELD STUDY EXPERIMENTAL DESIGN . . . . . . . . . . . . FORM FOR RECORDING DETECTION EVENTS . . . . . . . . . . QUARTERLY DWI ARREST RATES FOR ALL 10 PARTICIPATING AGENCIES DURING BASELINE AND TEST PERIODS . . . . . . . COMPARISON OF AVERAGE CUE PROBABILITY VALUES OF THE GUIDE WITH THOSE CALCULATED FROM THE FIELD TEST . . . . . . . . . . . . . . . . . . . . . DISPOSITION OF 4019 APPREHENSIONS RECORDED ON DRUNK DRIVER DETECTION GUIDE AND RECORD FORMS DURING THREE-MONTH TEST PERIOD
o

Page 14 17

22

29

33 41

MODIFIED DRUNK DRIVER DETECTION GUIDE

vii

LIST OF TABLES
Table 1 2 DWI Arrests per 100 Person-Hours of Patrol During Baseline and Test Periods by Agency . . . . . . . DWI Arrests per 100 Person-Hours of Patrol During Baseline and Test Periods by Patrol Type . . . . Mean Number of Cues Reported per DWI Arrest During Baseline and Test Periods by Agency . . . . . . . Frequency D i s t r i b u t i o n s of Detection Cues Recorded on DWI Arrest Reports During Baseline and Test Periods . . . . . . . . . . . . . . . . . . . . . 5 Mean BAC of Persons Arrested During Baseline and Test Periods by A g e n c y . . . . . . . . . . . . . . . Page 23 25 26

27 28

Comparison of Cue P r o b a b i l i t y Values of Detection Guide with Yalues Obtained from Field-Test Data for P(BAC > .10) . . . . . . . . . . . . . . . . Comparison of Cue P r o b a b i l i t y Values of Detection Guide with Values Obtained from F i e l d - T e s t Data for P(BAC ~ .05) . . . . . . . . . . . . . . . . .
8

30

31

Cue D ~ s c r i m i n a b i l i t y Values Computed from 4662 Detections Made During Detection and Field Studies: P(BAC ~ .I0) Cue D i s c r i m i n a b i l i t y Values Computed from 4662 Detections Made During Detection and Field Studies: P(BAC ~ .05) . . . . . . . . . . . , .....

35

36

10

Summary of the Content Analysis of Group Discussions Conducted at Each P a r t i c i p a t i n g Agency Regarding
Guide U t i l i t y ~ 37

vi ii

TECHNICAL SUMMARY
On-the-road detection of d r i v i n g cers. while intoxicated (DWI) involves offi-

the observation and i n t e r p r e t a t i o n of visual cues by police patrol

The effectiveness of DWI detection depends not only on the frequency

w i t h which patrol o f f i c e r s see and recognize cues i n d i c a t i v e of DWI, but on the extent to which observed cues d i s c r i m i n a t e between DWI and d r i v i n g while sober (DWS). This research project was conducted to answer the f o l l o w i n g questions: f o r DWI detection? What visual cues occur f r e q u e n t l y enough to b e u s e f u l To what extent do d i f f e r e n t cues d i s c r i m i n a t e between be

DWI and DWS? How can information on cue occurrence and d i s c r i m i n a b i l i t y used best f o r on-the-road detection of DWI? In the f i r s t phase of the p r o j e c t , reported e a r l i e r ,

the l i t e r a t u r e

was reviewed, DWI detection experts were interviewed, conducted to obtain data on the r e l a t i v e d i s c r i m i n a b i l i t y occurrence of visual detection cues. guide designed to f a c i l i t a t e the-road detection of DWI.

a large sample of and frequency of phase

arrest reports was analyzed, and an on-the-road study of DWI detection was The end product of the f i r s t

was a set of conclusions about DWI detection, and a prototype DWI detection a p p l i c a t i o n of the research f i n d i n g s to on-

In the second phase of the p r o j e c t ,

reported here, a DWI Detection The Guide was

Guide and an explanatory booklet were developed and tested. a small card of white p l a s t i c printed w i t h blue. The f i e l d

t e s t of the Guide was conducted with a sample of 10 law The t e s t analyses of of using the

enforcement agencies at locations throughout the United States. was designed to provide both l o n g i t u d i n a l and cross-sectional several measures l i k e l y to r e f l e c t the impact and u t i l i t y Guide, and to v e r i f y the values contained in the Guide.

L ,L

Use of the Guide was accompanied by a s t a t i s t i c a l l y overall increase in DWI a r r e s t rate of 12 percent.

significant

Rates during a 3-

month period, in which the Guide was used, were compared to a 12-month baseiine period. Five i n d i v i d u a l agencies had s i g n i f i c a n t increases of up to 94 percent; four agencies had no change; and one agency had a sign i f i c a n t decrease. Although there were no s t a t i s t i c a l l y s i g n i f i c a n t changes in de-

t e c t i o n practices r e f l e c t e d by greater use of the more d i s c r i m i n a t i n g cues or by arrests of drivers with lower BAC l e v e l s , trends were in those directions. The DWI p r o b a b i l i t y values contained in the Guide were v e r i f i e d by the f i e l d - t e s t results. Average ~alues obtained during the f i e l d w i t h corresponding t e s t were e s s e n t i a l l y the same as average values on the Guide, and Guide values f o r i n d i v i d u a l cues correlated s i g n i f i c a n t l y values calculated from f i e l d - t e s t data.

Experienced police o f f i c e r s who used the Guide expressed doubts about i t s value in improving t h e i r own DWI enforcement a b i l i t y . to important cues, t r a i n i n g new patrol o f f i c e r s , ports, and supporting court testimony. F i e l d - t e s t r e s u l t s led to a couple minor m o d i f i c a t i o n s in the Guide and the explanatory booklet. DWI detection p r o b a b i l i t i e s were based Several other minor the Guide. The on combined data from the early detection study and from the f i e l d t e s t , thus p r o v i d i n g a data base of 4662 detection events. m o d i f i c a t i o n s were made to f u r t h e r s i m p l i f y and c l a r i f y However, they considered the Guide to be valuable f o r increasing patrol s e n s i t i v i t y preparing DWI a r r e s t re-

r e s u l t i n g DWI Detection Guide is shown below in actual s i z e .

DWI DETECTION GUIDE

Chances in 100 of nighttime driver w i t h BAC equal o r greater than .10 -~ 65 65 GO GO GO 55 55 50 50 50 50 45 45 45 40 40 35 35 30 30

TURNING WITH WIDE RADIUS . . . . . . . . . . " ............... STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . . . . . . APPEARING TO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ALMOST STRIKING OBJECT OR VEHICLE . . . . . . . . . . . . . . . . . . WEAVING . . . . . . . . . . . . . DRIVING ON OTHER THAN DESIGNATED ROADWAY . . . . . . . . . SWERVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SLOW SPEED (MORETHAN 10 MPH BELOW LIMIT( . . . . . . . . STOPPING (WITHOUT CAUSE) IN TRAFFIC LANE . . . . . . . . . . FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DRIFTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TIRES ON CENTER OR LANE MARKER . . . . . . . . . . . . . . . . . . . . . BRAKING ERRATICALLY., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DRIVING INTO OPPOSING OR CROSSING TRAFFIC . . . . . . . . . . SIGNALLING INCONSISTENT WITH DRIVING ACTIONS . . . . . SLOW RESPONSE TO TRAFFIC SIGNALS . . . . . . . . . . . . . . . . . . STOPPING INAPPROPRIATELY |OTHER THAN IN LANE) . . . . TURNING ABRUPTLY OR I L L E G A L L Y . . . . . . . . . . . . . . . . . . . . . . ACCELERATINGOR DECELERATING RAPIDLY . . . . . . . . . . . . . . HEADLIGHTS OFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Special a d j u s t m e n t to the cue values 2 o r more cues observed: add 10 to the larger value BAG equal to o r greater t h a n .05: add 15 t o the value obtained for BAC equal to o r greater t h a n .10

A short 16-mm sound f i l m in color d e t e c t i o n cues and introduce the Guide

was produced

to

describe

the

,:

INTRODUCTION
On-the-road detection of d r i v i n g while intoxicated (DWI) involves the observation and i n t e r p r e t a t i o n of visual cues by police patrol o f f i cers. The effectiveness of DWI detection depends not only on the frequency with which patrol officers see and recognize cues i n d i c a t i v e of DWI, but on the extent to which observed cues discriminate between DWI and d r i v i n g while sober (DWS). This research project was conducted to answer the f o l l o w i n g three questions: What visual detection? cues occur frequently enough to be useful for DWI \

e To what extent do d i f f e r e n t cues discriminate between DWI and DWS? How can findings on cue occurrence and d i s c r i m i n a b i l i t y best be incorporated into practical procedures f o r on-the-road detection of DWI? The i n i t i a l phase of the project addressed and provided preliminary answers to these three questions by i d e n t i f y i n g useful visual cues and developing DWI detection methods. The primary products of the ~ i n i t i a l In this, phase were an interim report I and a proposed DWI detection guide.

-I

I I J /

f i n a l phase of the project, a Drunk Driver Detection Guide was d e v e l o p e d , l y and a f i e l d test was conducted to evaluate and v e r i f y the Guide. Prior t o | describing the objectives, methodology, and results of the f i e l d test, a summary of the interim report is provided as background.

1Harris, D. H., Howlett, J. B., and Ridgeway, R. G. The visual detection of d r i v i n ~ while intoxicated, project interim r e p o r t I d e n t i f i c a t i o n of visual cues and development of detection methods. Anacapa Sciences, Inc., f o r Department of Transportation, National Highway T r a f f i c Safety Administ r a t i o n , January 1979. (NHTSA Technical Reference No HS 805 051- NTIS No. PB 80 108 327). " '

BACKGROUND Only a very small proportion offense--only tion, about one in 2000. of persons DWI are arrested f o r t h i s Reasons f o r a low arrest rate might

include l i m i t a t i o n s on enforcement resources, lack of enforcement motivai n a b i l i t y to detect DWI, and others. However, previous research has also shown that even when persons DWI have been observed by police o f f i c e r s who were highly motivated to arrest f o r DWI, the arrest rate was r e l a t i v e l y

lOWo
As determined from roadside breathtesting surveys conducted throughout the United States, about six percent of drivers at night have a blood alcohol concentration (BAC) equal to or greater than .10. have a BAC equal to or greater than .05. About 15 percent Thus, i f DWI were defined at the Visual cues that are

BAC ~ .10 l e v e l , the p r o b a b i l i t y of detecting DWI from a random stop would be .06; at BAC ~ .05, the p r o b a b i l i t y would be .15. tion p r o b a b i l i t i e s ability DWI. above these chance levels. capable of d i s c r i m i n a t i n g between DWI and DWS can serve to increase detecThus, the key to enhanced discriminon-the-road detection of DWI is determination of the r e l a t i v e

of visual cues that are l i k e l y to be observed in association with

Previous Research

Many studies have investigated the effect of alcohol on driving behavior. They have employed laboratory apparatus, driving simulators, However, results have been only Although and instrumented vehicles in the f i e l d .

indirectly relevant to the objectives of the present project.

substantial evidence has been developed to indicate that alcohol-induced driver impairment is exhibited mainly in four driving functions--steering control, velocity control, time-sharing of attention, and information processing--the findings have not been specific enough to permit the identification and assessment of visual cues for on-the-road detection.

Lists

of cues have been developed through interviews with police developed from The r e s u l t i n g l i s t i n g s have been however, they have been of there can

o f f i c e r s experienced in DWI detection, including a l i s t i n g a survey conducted in the present study. only l i m i t e d use for DWI detection. both comprehensive and l o g i c a l l y organized;

Without information about the r e l a t i v e

frequencies of cue occurrence and r e l a t i v e cue d i s c r i m i n a b i l i t y ,

be no basis for defining useful visual cues or developing p r a c t i c a l guidelines f o r DWI detection. Analysis of DWI Arrests An analysis was completed of a sample of 1288 DWI arrest reports obtained from nine d i f f e r e n t police agencies throughout the United States f o r arrests made during the previous year. A t o t a l of 3658 visual detect i o n cues was reported in the sample, an average of about three cues per arrest. Frequency d i s t r i b u t i o n s prepared from the data, combined with the obtained from experienced of 129 visual cues potenr e s u l t s of previous research and cue l i s t i n g s patrol o f f i c e r s , provided a preliminary l i s t i n g t i a l l y useful f o r DWI detection.

On-the-Road Detection Study

An on-the-road study of D W I detection was conducted to determine the relative discriminability and frequency of occurrence of visual detection cues, under conditions typically encountered by police officers. driving behavior and, vehicle actions that deviated from normal. the driver with a portable breath tester. Trained In each observers accompanied police officers on patrol and recorded instances of instance, the police officer stopped the vehicle and measured the B A C of In addition to cue descriptions Since the data and B A C level, the observer recorded the circumstances and conditions under which the stop was made, and driver characteristics. collection effort required conducting pre-arrest breath tests of drivers, the study was conducted in two states, Indiana and North Carolina, that permitted, by statute, pre-arrest breath testing.

A t o t a l of 643 DWI detection events was observed and recorded. The sample was comparable to the national sample of 1288 DWI arrests in terms of time of day of stops, location (urban vs r u r a l ) of the stops, and sex of the d r i v e r . As expected, the main way in which the detection study sample of the BAC In the detection study, 39 percent of the drivers By contrast, 96 percent of the sample of DWI d i f f e r e d from the arrest report sample was in the d i s t r i b u t i o n levels of the drivers.

had a BAC ~ .05; 23 percent had a BAC in the range from .05 to 0.10; and 38 percent had a BAC ~ .10. arrests reported drivers with BAC ~ .10. Analyses of the 1681 cue occurrences recorded during the 643 detect i o n events included: computation of cue frequencies, c a l c u l a t i o n of cue d i s c r i m a b i l i t y values, study of cue co-occurrence, assessment of cue order of appearance, and c o r r e l a t i o n a l ditions. redefined, As part of the analyses to determine the impact on cue characteristics, and conand effort, cues were recombined study. occurrence of a l t e r n a t i v e detection s t r a t e g i e s , analytical ultimately,

into a set of 23 visual cues that accounted f o r 93 detection The f o l l o w i n g

percent of the cue occurrences in the

conclusions were developed from the r e s u l t s of the study: Although the potential number of visual detection cues is very large, most detection events can be accounted for by a r e l a t i v e l y small number of cues. T y p i c a l l y , a detection cue is observed with one or more other cues; however, there are few subsets of s p e c i f i c cues that occur f r e quently together. There are large differences among visual detection cues in the frequency with which they occur with DWI, and in t h e i r a b i l i t y to discriminate between DWI and DWS. In general, the conditions ( l i g h t i n g , time of day, distance, locat i o n , vehicle condition, type of roadway, age or sex of d r i v e r ) under whichcues are observed have r e l a t i v e l y l i t t l e influence on cue occurrence.

 Patrol strategy (general patrol vs. patrol with DWI emphasis) g r e a t l y affects the r e l a t i v e frequencies with which cues are observed. DWI Detection Guide A p r e l i m i n a r y DWI d e t e c t i o n patrol officers. guide was developed to f a c i l i t a t e the

a p p | i c a t i o n of research f i n d i n g s to on-the-road detection of DWI by p o l i c e The extent of competing demands placed upon patrol o f f i likely to be encountered, rapid response, the s t r i n g e n t and the large time, the need f o r c e r s - - t h e v a r i e t y of s i t u a t i o n s demands on available

amount of other law enforcement information that must also be learned and retained--suggested that the f i n d i n g s of t h i s simply and d i r e c t l y . research f i n d i n g s into a practical aid f o r study be presented f o r use DWI detection. Because the Guide developTherefore, the guide was developed to transform the

empirical r e s u l t s were not necessarily simple or free of s u b t l e t y , extrapo l a t i o n and judgment were exercised during t h i s process. ment was governed by the f o l l o w i n g c r i t e r i a : Account f o r the largest number of smallest number of detection cues. detection

events

with

the

Enhance the d i s c r i m i n a b i l i t y of a v a i l a b l e detection cues. Employ a p r o b a b i l i s t i c output. Accommodate m u l t i p l e cue occurrences. m Accommodate a l t e r n a t i v e enforcement statutes and p o l i c i e s . Emphasize s i m p l i c i t y , A DWI detection phase of the p r o j e c t . practicality, and ease of use. in t h i s initial

guide was developed conceptually

The concept was r e f i n e d and transformed during the into the Drunk Driver Detection Guide shown in

firstpart of the f i e l d - t e s t the Appendix.

FIELD-IEST OBJECTIVES
The purpose of the f i e l d test was to evaluate and verify-the Drunk Driver Detection Guide. There were four different facets of the f i e l d test, as reflected in the following four specific objectives: 1. Determine the potential impact of the Guide on DWI arrest rates.

The ultimate criterion for evaluating the Guide was the extent to which i t enhanced DWI enforcement, as reflected by increased DWI arrest rates. However, although detection accuracy is l i k e l y to influence enforcement rate, it is surely not the only factor that does. and, second, to arrest Other factors also influence the decision of the patrol o f f i c e r , f i r s t , apprehend a driver driver. to apprehend or not

or release the apprehended

As a consequence, the ultimate c r i t e r i o n of enhanced DWI enforceis l i k e l y to be contaminated for On the other hand, i f use of the Guide

ment, as measured by DWI arrest rate, purposes of evaluatingthe Guide.

cannot be shown, under present circumstances, to have some positive impact on the rate of DWI arrests, its ultimate contribution to DWI enforcement w i l l probably be minimal.
Lo

Determine the extent to which DWI detection practices are Is the form of the Guide and the

changed through use of the Guide.

training provided for its use adequate to modify DWI detection practices? Are the.more discriminating cues reported more frequently as a result of using the Guide? Are average B A C levels of arrested drivers DWI lower as a result of detecting and arresting more drivers near the .10 threshold? 3. Verify the Guide. Development of the Guide was based mainly on Prior to any widespread

data collected during 643 DWI detection events.

use of the Guide, cue frequency distributions and d i s c r i m i n a b i l i t y values require verification through additional data obtained from additional DWI detection events. 4. Evaluate the Guide as a practical, useful detection aid. benefit? Is

the Guide too awkward to be of practical

Is the Guide too

10

simple to provide the information little face v a l i d i t y to i n s t i l l

needed?

Does the Guide provide too These and other information that

confidence in the user? study to obtain

questions were addressed in the f i e l d might be useful to modify the Guide. The f i e l d of DWI. patro! officers

test was limited to an assessment of the visual detection detections were r e s t r i c t e d to those made v i s u a l l y by suspected of DWI was in the car. DWI while the driver

Therefore,

arrests made as a consequence of an accident or by an o f f i c e r dispatched in response to a request were eliminated.

11

METHOD
The f i e l d test of the Guide was conducted with a sample of 10 law enforcement agencies at locations throughout the United States. The test was designed to provide both longitudinal and cross-sectional analyses of several measures l i k e l y to r e f l e c t the impact and u t i l i t y of using the Drunk Driver Detection Guide, and to v e r i f y the detection p r o b a b i l i t i e s contained in the Guide.

FIELD-TEST DESIGN
The design was a compromise between what would be required for the most d e f i n i t i v e possible evaluation and what was p r a c t i c a l l y feasible in terms of police cooperation and available resources. Under the pressures procedures individual that existed throughout the country for increased police e f f i c i e n c i e s and f o r more police s e n s i t i v i t y to individual privacy, f i e l d - t e s t could be neither burdensome nor p o t e n t i a l l y embarrassing to police agencies. f i e l d test was designed.

Consequently, i t was w i t h i n these constraints that the Although less than optimal from a t h e o r e t i c a l

perspective, the design did provide an adequate basis for evaluation of the Guide while placing a minimum burden on the part of p a r t i c i p a t i n g police agencies.

Experimental Design
A within-subjects type of experimental design was employed with 10 p a r t i c i p a t i n g police agencies. Measures related to DWI enforcement effectiveness were obtained from each agency during a 12-month baseline period and during a three-month test period in which the Guide was used. measures were obtained during both baseline and test periods: rate, arrested. Three DWI arrest

frequencies of reported detection cues, and BAC levels of persons During the three-month test period, two additional measures r a t i o of drivers DWI to drivers apprehended for each cue or The experimental design is i l l u s t r a t e d

were obtained:

cue combination, and opinions and suggestions of p a r t i c i p a t i n g police off i c e r s regarding use of the Guide. in the diagram of Figure 1.

12

Field-Test Measures
The f i v e d i f f e r e n t measures obtained during the f i e l d d i r e c t l y to one or more of the f i e l d - t e s t described objectives. in the f o l l o w i n g paragraphs; the d a t a - c o l l e c t i o n test related and analysis

The measures are

procedures required by each measure are discussed l a t e r . DWI arrest r a t e . This measure was defined as the number of DWI Rates were calculated monthly for

arrests made per unit of patrol e f f o r t .

each agency, in terms of number of DWI arrests per 100 person-hours of patrol a c t i v i t y , during both baseline and test periods.

Frequencies of reported detection cues. Frequency distributions o f

cues reported on DWI arrest reports were obtained during both baseline and test periods. Comparisons of these distributions might reveal whether or not use of the Guide resulted in any changes in detection cues employed. BAC levels. Measured B A C levels of persons arrested for DWI during Enhanced detection of persons DWI might be

the period were obtained for comparison with BAC levels of persons arrested during the baseline period. r e f l e c t e d in decreased BAC levels, as officers become more f a c i l e in detecting BAC levels closer to the legal impairment level. DWI detection probabilities. outcome of each detection event. For purposes of the f i e l d test, the

Guide was modified to permit direct recording of observed cues and the These data provided the basis for calculating DWI detection probabilities associated with each cue or cue combination, f o r comparison to Guide values. Police o f f i c e r opinions and suggestions. test period. Opinions and suggestions

were obtained from groups of officers about midway through the three-month Responses were obtained by means of group interviews conducted by project staff members during agency v i s i t s .

13

S A M P L E OF P O L I C E A G E N C I E S

AIBICIDIEIFIGIHI

IIJ
m

DWl A R R E S T RATES CUE DISTRIBUTIONS BAC DISTRIBUTIONS

12 M O N T H S

I N I T I A T E USE OF T H E DWI D E T E C T I O N

ouo,

DWl A R R E S T RATES CUE D I S T R I B U T I O N S BAC DISTRIBUTIONS DWI D E T E C T I O N D A T A OPINIONS & SUGGESTIONS

3 MONTHS

Figure 1. Field study e x p e r i m e n t a l design.

14

PARTICIPATING AGENCIES

The 10 law enforcement agencies selected for participation as estimated by the 1970 census order: Agencies are listed

in the

study are listed below, along with the populations served by each agency, in alphabetical

Albuquerque (New Mexico) Police Department Englewood (Colorado) Police Department Eugene (Oregon) Police Department Evansville (Indiana) Police Department Monroe County (New York) S h e r i f f ' s Office Pulaski County (Arkansas) S h e r i f f ' s Office Santa Ana (California) Police Department Tacoma (Washington) Police Department Topeka (Kansas) Police Department Vanderburgh County (Indiana) S h e r i f f ' s Office

243,751 33,695 76,346 138,764 711,917 287,189 156,601 154,581 125,011 168,772

Within the 10 agencies, data were collected from different types of police patrols: nine agencies employed general patrols responsible for criminal and t r a f f i c enforcement and/or t r a f f i c patrols responsible for t r a f f i c enforcement; two agencies had DWI patrols responsible primarily for D~I enforcement; and one agency had a selective t r a f f i c patrol responsible for DWI and speeding enforcement. participated.
FIELD-TEST PROCEDURES

A total

of 466 patrol officers

The f i e l d test consisted of seven major tasks conducted sequentially. Each task is summarized and discussed in the following paragraphs.

15

!'

1.

Preparation of Guide and Materials

Alternative formats for the Drunk Driver Detection Guide were devel-

oped and exposed to samples of police officers who had participated earlier in the research. After consideration of the opinions and suggestions To accompany and An Explanation of The booklet received, f i n a l specifications for the Guide were prepared. The resulting Guide is described and i l l u s t r a t e d in the Appendix. explain the Guide a booklet, "Drunk Driver Detection:

the Drunk Driver Detection Guide," was prepared and printed. is also presented in the Appendix.

For collecting information for Guide v e r i f i c a t i o n , a special record form was designed for use during the three-month test period. levels, and indicate the disposition of the apprehended driver. This form Sets of 25 enabled the police officer to check the cues observed, record estimated BAC forms were combined into a 10.16 cm x 20.96 cm (4 inches x 8~ inches) pad. The form was about the size of most c i t a t i o n books to f a c i l i t a t e handling. As a form was completed and removed for submittal to Anacapa, a new form was exposed for recording the next detection event. trated in Figure 2. The form is i l l u s -

2.

Arrangementof Agency Participation

Ten law enforcement agencies were selected for participation in the

study.

A l i s t of the participating agencies along with the size populaAgency selection was made in

tions they served was presented earlier. accordance with the following c r i t e r i a .

e Geographical dispersion throughout the United States, agencies from the West, North, East, South and Central regions of the country. Agency interest in enforcing statutes which prohibit DWI, although special DWI patrols or practices were not required. Agency willingness to cooperate in accordance with the requirements of the study.

16

I
I I l I I

DO NOT WRITE IN THIS SPACE

I I - - I

i:il ..

REPORT NO. OFFICER ID:

i .

ESTIMATED BAC [ ] BELOW 0.05 OF THE DRIVER: [ ] 0 . 0 5 t o 0 . 1 0 [ ] 0.10 AND ABOVE DISPOSITION: DAY__ YEAR__ [ ] RELEASED [ ] ARRESTED

MONTH__ Check appropriate box(es)

Percentage of nighttime drivers with BAC equal to or greater than .10 ].. Visual Cues

(01) [ ] (02) [ ] (o3) [ ] (o4) [ ] (05) [ ] (06) [ ] (07) [ ] (co) [ ] (09) [ ] (10) [ ] (11)[] (t2) [ ] (13) [ ] (14)[] 05) [ ] (16) [ ] (17) [ ] (18) [] (19) [ ] (20) [ ] (21) [ ] (22) [ ] (23) [ ] (24) [ ]

STOPPING [WITHOUT CAUSE] IN TRAFFIC LANE . . . . . . 70 FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . 60 TURNING WITH WIDE RADIUS . . . . . . . . . . . . . . . . . . . . . 60 APPEARING TO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . 60 DRIVING ON OTHER THAN DESIGNATED ROADWAY . . . . 55 STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . 55 ALMOST STRIKING OBJECT OR VEHICLE . . . . . . . . . . . . 55 SLOW RESPONSE TO TRAFFIC SIGNALS . . . . . . . . . . . . . 50 HEADLIGHTS OFF [AT NIGHT] . . . . . . . . . . . . . . . . . . . . . 50 SIGNALLINGINCONSISTENT WITH DRIVING ACTIONS .. 45 WEAVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 TIRES ON CENTER OR LANE MARKER . . . . . . . . . . . . . . . 45 DRIFTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 SWERVING " 45 ACCELERATING OR DECELERATING RAPIDLY . . . . . . . . . 45 SLOW SPEED [MORE THAN 10 MPH BELOW LIMIT] . . . . 45 FAST SPEED [MORE THAN 10 MPH ABOVE LIMIT ] . . . . . 35 FAILING TO RESPOND TO TRAFFIC SIGNALS OR SIGNS . 35 BRAKING ERRATICALLY 35 STOPPING INAPPROPRIATELY [OTHER THAN IN LANE] , 35 TURNING ABRUPTLY OR ILLEGALLY . . . . . . . . . . . . . . . . 30 DRIVING INTO OPPOSING OR CROSSING TRAFFIC . . . . . 30 DRIVING WITH VEHICLE DEFECT[S] . . . . . . . . . . . . . . . . 30 OTHER:

~.:.. , :::: !:i:-i : :i;i:;i :::;. ii ;.iii ~:i~i.i: :~.

~-i:, ~:i::,: i~:~i~i:il ;::!ii : i~.ii;ii:

.........

..... Special Adjustments to the Percentages ;:~i:;i" 2 cues: Add 5 to the larger percentage. :!:i ::: - 3 o r more cues: Add 10 to the largest percentage, !~i;': To predict BAC equal to or greater than .05: Add 20 to the per:fill. centage.10, obtained for drivers with BAC equal to or greater than

::i: i!:; ;.:.i, ;:;:: ~!:~i':~::::

Figure 2. Form for recording detection events.

17

Formal contact was made i n i t i a l l y

by mail with f o u r agencies in each Visits The

NHTSA region, followed by telephone contact with those responding. final

were then made to the 13 agencies t h a t appeared to be most promising. regarding agency i n t e r e s t fic 3. and w i l l i n g n e s s . The e f f o r t

i0 were selected on the basis of what was learned during the v i s i t s was c l o s e l y coor-

dinated with the Contract Technical Manager, cognizant personnel of TrafSafety Programs, and NHTSA Regional Coordinators. Collection of Baseline Data Three types of baseline data were collected from each agency--number of DWI arrests, frequencies of reported cues, and BAC levels of arrested drivers. reviewed, All reports of DWI arrests over the 12-month baseline period were and the required data abstracted from them. In addition, a

month-by-month tabulation of the number of person-hours of patrol a c t i v i t y was recorded. This information was used to calculate DWI arrest rate, number of DWI arrests per 100 person-hours of patrol a c t i v i t y . 4. Initiation of Guide Use During the month p r i o r to the three-month t e s t period, t r a i n i n g was conducted to introduce the Guide to the 466 p a r t i c i p a t i n g p a t r o l o f f i c e r s w i t h i n the 10 agencies. The t r a i n i n g required about 30 minutes and, in or r o u t i n e briefing sessions. self-explanatory, the t r a i n i n g many cases was conducted during r o l l - c a l l Although the Guide i t s e l f was r e l a t i v e l y sessions provided an opportunity to:

Explain the empirical basis f o r the Guide, search that went into i t s development.

summarizing the re-

Assure t h a t the three-step d e t e c t i o n process and the underlying concepts of p r o b a b i l i t y were understood. ~rovide an opportunity to c l a r i f y d e f i n i t i o n s of the visual detect i o n cues and d i s t i n c t i o n s among them. Answer questions concerning the Guide and i t s use.

18

 Explain the purpose of recording detection events and the use of the Drunk Driver Detection Guide and Record Form including ins t r u c t i o n s for submitting completed sheets to Anacapa. 5. Data Collection During the three-month field-test and data-collection period,

a r r e s t - r a t e , cue-frequency, and BAC-level data were collected in the same manner as the baseline data were collected e a r l i e r . DWI detection-event data were collected by means of the detection and shown in Figure 1. These forms were one form was completed by an Anacapa

report forms discussed e a r l i e r f o r each detection event.

kept in the patrol vehicle by each patrol o f f i c e r ;

Completed forms were then given to a designated

member of the agency who either held them for c o l l e c t i o n s t a f f member or forwarded them by mail to Anacapa.

Police opinions and suggestions concerning the use of the Guide were obtained during group discussions, period. any topic related to the Guide. minimum: about midway through the three-month Opportunity and encouragement were provided f o r the group to cover Topics covered in each session were, ata

o How useful has the Guide been to you in the enforcement of DWI? e S p e c i f i c a l l y , what advantages and disadvantages have you found in using the guide? o What specific suggestions do you have for improving the Guide and increasing i t s usefulness? During the data-collection period, project staff members v i s i t e d

each agency about three times to assure that data were being collected properly.

6.

Data Analysis
Consistent with the four f i e l d - t e s t objectives and the types of data

c o l l e c t e d , the data-analysis e f f o r t consisted of the f o l l o w i n g four parts:

19

Comparisons of DWI arrest rates. Arrest rates (number of D W I arrests per 100 person-hours of patrol) were calculated monthly for each of
the ten police agencies, and patrol types within agencies, for each of the 12 months prior to and for each of the three months following i n i t i a t i o n of the use of the Guide. These data were charted and comparisons made between baseline and test periods. and for types of patrols. Baseline-test comparisons were made for indiStatistical tests were conducted using both Comparisons were made between Statisvidual agencies and patrols for the total sample of agencies and patrols, analysis of variance (ANOVA) and X2.

Comparisons of cues and B A C levels.

baseline and test periods for:

number of cues employed, frequency d i s t r i -

butions of cues reported, and mean B A C levels of drivers arrested.

t i c a l tests were made using 2 and t sampling distributions to determine the significance of any differences found in these variables between baseline and test periods. Guide verification. D W I probability values were calculated from

f i e l d data for each cue, under each cue occurrence p o s s i b i l i t y (one of one or more cues, one of two or more cues, or one of three or more cues), and compared with D W I probability values obtained from the Guide.

Tabulation of patrol officer opinions and suggestions.

garding u t i l i t y of the Guide. each opinion or suggestion made.

A content

analysis was completed of responses recorded from group discussions reResponses were classified and tabulated; notation was made regarding whether or not there was group consensus for

7.

Preparation of Final Report

The findings resulting from the four components of the f i e l d - t e s t

data collection and analysis effort were integrated and interpreted in terms Guide. report. of their implication for the modification and future use of the The study was then described and the results presented in this

20

RESULTS
Use of the Drunk Driver Detection Guide was accompanied by a s i g n i f i cant increase in DWI arrest rate. However, use of the Guide was not accompanied by increases in the number of cues reported per DWI arrest, changes in cue emphasis, or reductions in BAC levels of drivers arrested. The p r o b a b i l i t y field-test results. cantly different values contained in the Guide were v e r i f i e d by

Average Guide and f i e l d - t e s t

values were not s i g n i f i Field-test

for either P(BAC ~ .10) or P(BAC ~ .05).

p r o b a b i l i t y values for individual cues correlated s i g n i f i c a n t l y with Guide p r o b a b i l i t y values. Police o f f i c e r s who used the Guide expressed doubts about i t s value f o r increasing t h e i r own a b i l i t y to detect DWIo However, they considered the Guide to be a valuable aid f o r increasing patrol cues, t r a i n i n g inexperienced patrol o f f i c e r s , reports, and supporting court testimony.
IMPACT O N DWI ARRESTRATES

awareness of useful DWI arrest

preparing

C o l l e c t i v e l y , for all 10 p a r t i c i p a t i n g agencies, DWI arrest rate was 12 percent higher during the test period than during the baseline period. This difference was s t a t i s t i c a l l y five significant (p < .01). Comparison of Individually, had rates that baseline and t e s t periods is shown g r a p h i c a l l y in Figure 3. agencies had DWI arrest rates s i g n i f i c a n t l y lower rate; period; one had a s i g n i f i c a n t l y remained e s s e n t i a l l y unchanged. and four

higher during the t e s t

Arrest rate was defined and calculated as the number of DWI arrests made per I00 person-hours of patrol a c t i v i t y . hours of p a t r o l . During the e n t i r e 15-month period, a t o t a l of 5348 arrests were made during a t o t a l of 788,200 personArrest rates are provided in Table 1 for each agency, f o r and person-hours of patrol activity from which they the 12-month baseline period and three-month test period, along with the numbers of arrests

21

1.0

0.9
.-I

O <c
It.

0.8

O3

0.7

O
Z

0.6

O3 IJJ

0.5

"" 1:
U.I a. 03

0.4

P"
03 U.I

0.3

<~

0.2

0.1

AUG-OCT 1978

NOV-JAN

FEB-APR

MAY-JUL 1979 I

AUG-OCT 1979 TEST I

BASELINE Figure 3. Quarterly DWl arrest rates for all ten participating agencies during baseline and test periods.

22

TABLE 1 DWI Arrests per 100 Person-Hours of Patrol During Baseline and Test Periods by Agency BASELINE PERIOD TEST PERIOD 2 38.68* 34.44* 12.17" 30.83* .31 .88 .33 48.91" .63 10.50" 10.57"

DWI PATROL ARREST % DWI PATROL ARREST AGENCY ARRESTS I00 HRS, RATE ARRESTS I00 HRS. RATE CHANGE A
B

2017 319 150 234 66 148 65 530 130 526 4185

1387 641 639 1346 269 728 102 68 362 758 6300

1.45 .50 .23 .17 .25 .20 .64 7.79 .36 .69 .66

379 135 67 109 14 30 13 217 28 171 1163

369 150 171 331 65 176 25 16 93 186 1582

1.03 .90 .39 .33 .22 .17 .52 13.56 .30 .92 .74

-29 +80 +70 +94 -12 -15 -19 +74 -17 +33 +12

C
D E

F G H I J TOTAL *p < .01

were calculated.

The s t a t i s t i c a l

s i g n i f i c a n c e of d i f f e r e n c e s in arrest by a randomly assigned

rates between baseline and test periods were tested by means of the 2 sampl~ng d i s t r i b u t i o n . Agencies are i d e n t i f i e d l e t t e r f o r presentation of the r e s u l t s . Some of the differences types of patrols for included the full in

among agencies in magnitude of DWI arrest The the of field study were: and general patrols

rates were a function of the types of enforcement p a t r o l s employed. responsible range criminal traffic

enforcement

23

activities, and a

traffic

patrols traffic

responsible patrol

for

the f u l l for

range of t r a f f i c DWI and speeding

enforcement, DWI patrols responsible mainly f o r DWI detection and a r r e s t , selective patrols responsible enforcement. traffic Within nine agencies data were collected from general and/or (GENERAL); w i t h i n two agencies data were collected from (SELECTIVE). Agency A had general and DWI p a t r o l s ; and data were c o l l e c t e d only

DWI patrols (DWI); and w i t h i n one agency data were collected from a selective traffic patrol Agency J had general and selective p a t r o l s ; from DWI p a t r o l s in Agency H.

In Agencies A and J, the percentage changes

in DWI arrest rate from baseline to t e s t periods were comparable between the two patrol types; thus, data from both types of patrol were combined for these two agencies in Table 1. Arrest rates are shown in Table 2 f o r each type of patrol general and selective t r a f f i c patrol decrease in arrest rate, patrols, but not f o r DWI p a t r o l s . significant There The DWI of

were s i g n i f i c a n t increases in arrest rate from baseline to t e s t periods f o r of Agency A had a s t a t i s t i c a l l y (p < .01) 27 percent

from 18.50 to

13.50 while the DWI patrol

Agency H had a s t a t i s t i c a l l y 7.79 to 13.56.

s i g n i f i c a n t (p < . 0 1 ) 74 percent increase from signifi-

The decrease in one cancelled the increase in the other,

r e s u l t i n g in a small increase overall that was not s t a t i s t i c a l l y cant. Striking differences among arrest rates for the three

types

of

patrols are also r e f l e c t e d in Table 2.

Arrest rate of s e l e c t i v e t r a f f i c arrest rate of DWI these data and eight times that

p a t r o l s was about four times that of general p a t r o l s ; p a t r o l s was about 30 times that of general p a t r o l s , of the s e l e c t i v e traffic patrol.

F r o m another perspective,

revealed that one DWI arrest was made f o r each eight hours of DWI p a t r o l , 70 hours of selective t r a f f i c patrol, or 260 hours of general p a t r o l .

CHANGES IN DWI DETECTION PRACTICES

There were no s t a t i s t i c a l l y significant differences between baseline and test periods on any of three measures that might have reflected

24

TABLE 2 DWI Arrests per 100 Person-Hours of Patrol During Baseline and Test Periods by Patrol Type BASELINE PERIOD PATROL TYPE GENERAL SELECTIVE DWI TOTAL *p < .05 * * p < .01 TEST PERIOD

DWI PATROL ARREST DWI PATROL ARRESTS 100 HRS. RATE ARRESTS 100 HRS. 2277 299 1582 4158 5967 208 125 6300 .38 1.44 12.66 .66 636 94 433 1163 1504 46 32 1582

% ARREST RATE CHANGE

.42 2.04 13.53 .74 +16 +42 +07 +12

2 5.34* 31.60"* 1.48 10.57"*

changes in DWI detection practices. recorded per arrest did not increase;

The number of visual a significant shift

detection cues to the use of of persons

higher p r o b a b i l i t y cues did not occur; arrested did not decrease s i g n i f i c a n t l y . As shown in Table 3, periods.

and the BAC levels

about 2.0 cues were recorded on DWI arrest were found among individual agen-

reports, on the average, over the 10 agencies during both baseline and test Although larger differences cies, most of which were increases, none was s t a t i s t i c a l l y p < .05. s i g n i f i c a n t at

The r e l a t i v e frequencies with which detection cues were reported on DWI arrest reports were very stable and r e l a t i v e l y unaffected by use of the Guide. Table 4 shows the frequency d i s t r i b u t i o n s of detection cues recorded during both baseline and test periods. To permit d i r e c t comparison,

the d i s t r i b u t i o n s shown were based on the number of occurrences of each cue f o r 1000 t o t a l cue occurrences. Inspection o f the two d i s t r i b u t i o n s reveals that they are nearly i d e n t i c a l .

25

TABLE 3 Mean Number of Cues Reported per DWI Arrest During Baseline and Test Periods by Agency BASELINE PERIOD AGENCY A
B

TEST PERIOD NUMBER MEAN ARRESTS CUES 379 135 67 109

14 % CHANGE

NUMBER ARRESTS 2017 319 150 234 66 148 65 530 130 526 4185

MEAN CUES 2.35 2.01 1.38 1.78 1.47 2.43 2.56 1.92 1.83 1.19 2.03

2.41 2.16 1.55 1.74 1.33 3.37 2.75 1.78 1.90 1.26 1.99

+02 +07 +12 -02 -09 +39 +07 -07 +04 +06 -02 sig-

C
D

E F G H
I

30 13 217 28 171 1163

J TOTAL

The product-moment c o r r e l a t i o n between the two is .98, s t a t i s t i c a l l y nificant beyond p < .01. There was some s h i f t

in frequencies toward the occurred

higher p r o b a b i l i t y cues during the t e s t period but t h i s was not s t a t i s t i cally significant. The 10 cues with highest DWI p r o b a b i l i t i e s 8 percent more f r e q u e n t l y in the t e s t period while the 10 cues with lowest DWI p r o b a b i l i t y occurred 9 percent less f r e q u e n t l y . Although the mean BAC of persons arrested f o r DWI decreased from the baseline period to the test period in 9 of 10 agencies, the amount of the decrease was not s t a t i s t i c a l l y in Table 5. significant. These r e s u l t s are presented

VERIFICATION OF THE GUIDE

P r o b a b i l i t y values contained in the Guide were v e r i f i e d by the proportions of drivers apprehended during the f i e l d test who were found

26

TABLE 4 Frequency D i s t r i b u t i o n s of Detection Cues Recorded"on DWI Arrest Reports During Baseline and Test Periods OCCURRENCE (TIMES IN 1000) BASELINE TEST lane 17 5 23 21 32 62 62 7 22 17 145 4 31 49 57 40 129 73 9 25 60 52 17 13 5 25 21 29 61 72 10 29 23 156 5 24 35 41 32 130 83 10 23 37 56 21

DETECTION CUE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Stopping (without cause) in t r a f f i c Following too closely Turning with wide radius Appearing to be drunk Driving on other than designated roadway Straddling center or lane marker Almost s t r i k i n g object or vehicle Slow response to t r a f f i c Headlights o f f (at night) S i g n a l l i n g inconsistent with d r i v i n g actions Weaving Tires on center or lane marker Drifting Swerving Accelerating or decelerating r a p i d l y Slow speed--more than 10 MPH below l i m i t Fast speed--more than 10 MPH above l i m i t F a i l i n g to respond to t r a f f i c Braking e r r a t i c a l l y signals or signs signals

Stopping inappropriately other than in lane Turning abruptly or i l l e g a l l y Driving into opposing or crossing t r a f f i c Driving with vehicle defect(s)

27

TABLE 5 Mean BAC of Persons Arrested During Baseline and Test Periods by Agency BASELINE PERIOD AGENCY A
B

TEST PERIOD NUMBER ARRESTS 313 88 57 89 12 19 11 210 19 153 971" MEAN BAC 162 .175
.168 %

NUMBER ARRESTS 1820 210 132 177 52 80 65 53O 84 475 3625*

MEAN BAC .163 .181 .182 .189 .193 .178 .181 .174 .160 .154 .167

CHANGE -01 -04 -08 +02 -03 -03 -06 -04 -14 -06 -02

C D E F G H I J TOTAL

192

187

171 170

167 144

137

164

* D i f f e r s from t o t a l numbers of a r r e s t s not obtained f o r every person arrested

because BAC was Average Guide

to have BAC levels of equal to or greater than .05 and .10. probability probabilities calculated from f i e l d - t e s t data.

values over a l l cues were e s s e n t i a l l y the same as the o v e r a l l Figure 4 shows comparisons

of Guide and t e s t values f o r P(BAC ~ .05) and P(BAC ~ . 1 0 ) when the observed cue was one of one or more cues, one of two or more cues, and one of three or more cues. C o r r e l a t i o n s between Guide DWI p r o b a b i l i t i e s ties for individual cases. 7. cues were s t a t i s t i c a l l y values and t e s t DWI p r o b a b i l i (p < .05) in a l l light of two

significant

Comparisons of Guide and t e s t values are presented in Tables 6 and must be made in First, the p r o b a b i l i t i e s contained in the Guide

Comparison of Guide and t e s t

important considerations

28

1.0

0.8

U3

0.6 ="" ~ ~ ~ TEST 0.4

AI (J ,

D o.

0.2

I
ONE + CUES

I
TWO + CUES

T
THREE+ CUES

1.0-

o~8
TEST-~ 0.6

^i

C) m n

0.4

0.2

ONE + CUES

TWO + CUES

THREE+ CUES

Figure 4. Comparison of average cue probability values of the Guide with those calculated from the field test.

29

TABLE 6 Comparison of Cue P r o b a b i l i t y Values of Detection Guide with Values Obtained from Field-Test Data f o r P(BAC ~ .10) P(BAC ~ .10) WHEN CUE OBSERVED AS ONE OF: TWO OR THREE OR ONE OR MORE CUES MORE CUES MORE CUES DETECTION CUE 1 2 3 4 5
" 6

GUIDE TEST GUIDE TEST GUIDE TEST lane

.70 .60 .60 .60
55

Stopping (without c a u s e ) i n t r a f f i c Following too closely Turning with wide radius Appearing to be drunk

.42 .38 . 64 .61

.56

.75
.65 65

.75
.44 .64

.80
.70

.83
.53

.70 .70
65

.72 .75
.76

.65 .60 .60 .60

55

.65 .64 .67 .71

.64

Driving on other than designated roadway Straddling center or lane marker Almost s t r i k i n g object or vehicle Slow response to t r a f f i c Headlights o f f (at night) S i g n a l l i n g i n c o n s i s t e n t with d r i v i n g actions Weaving Tires on center or lane marker Drifting Swerving Accelerating or decelerating r a p i d l y Slow speed (more than i0 MPH below l i m i t ) Fast speed (more than 10 MPH above l i m i t ) F a i l i n g to respond to t r a f f i c Braking e r r a t i c a l l y Stopping i n a p p r o p r i a t e l y other than in lane Turning abruptly or i l l e g a l l y Driving into opposing or crossing t r a f f i c Driving with vehicle defect(s) AVERAGE C o e f f i c i e n t of correlation~between Guide and test signals or sign signals

.55 .55 .50

50

.64 .65 .38

.24

.65
65

.68
.73

7 8 9 10 11 12 13 14 15 16 17 18
_

.60

.66

.55
.50

.41
.58

.60 .55
.55

.54 . 73
.76

.45 .45 .45

45

.37 .62 .46

.50

.50
.50 .50

.67
.47 .54

.55
55 .55

.53
.65 .66

.45 .45 .45 .35 .35 .35 .35 .30

30 30

.57 .26 .52 .13 .15 .52 .41 .34

.50 .06

.50 .50 .50 .40 .40

40

.61 .34 .65 .49 .39

.56

.55
.55

.52
.70

.45
45 .45

.59
.60 .59

19 20 21 22 23

.40 .35
35

. 52 .53
.58

.45 .40 .40

.40

.67 .54 .67

.44

.35 .51

.21 .55

.46

.43

.56

.65

.48

.54

.54

30

TABLE 7 Comparison of Cue P r o b a b i l i t y Values of Detection Guide with Values Obtained from F i e l d - T e s t Data for P(BAC ~ .05) P(BAC ~ . 0 5 ) WHEN CUE OBSERVED AS ONE OF: ONE OR TWO OR THREE OR MORE CUES MORE CUES MORE CUES GUIDE TEST GUIDE TEST GUIDE TEST lane .90 .80 .80 .80 .75 .75
75

DETECTION CUE 1 2 3 4 5
, 6

Stopping (without cause) in t r a f f i c Following too closely Turning with wide radius Appearing to be drunk

.51 .62 .79 .76 .67 .78

.76

.95 .85 .85 .85 .80 .80 .80 .75 .75 .70 .70 .70 .70 .70 .70 .70 ,60 .60 .60 .60 .55 .55 .55 .71

.80 .63 .81 .76 .77 .79 .81 .69 .54 .62 .8O .61 .73 .71 .47 .80 .66 .55 .72 .57 .65 .72 .38 .68
i

1.00 ,90 .90 .90 .85 .85 .85 .80 .80 .75 .75 .75 .75 .75 .75 .75 .65 .65 .65 .65 .60 .60 .60 .76 .59

.92 .71 .86 .82 .88 .80 .82 .68 .69 .82 .85 .63 .81 .76 .62 .78 .74 .72 .67 .67 .70 .78 .58 .75

Driving on other than designated roadway Straddling center or lane marker Almost s t r i k i n g object or vehicle Slow response to t r a f f i c s i g n a l s Headlights o f f (at night) S i g n a l l i n g inconsistent with driving actions Weaving Tires on center or lane marker Drifting Swerving Accelerating or decelerating r a p i d l y Slow speed (more than 10 MPH below l i m i t ) Fast speed (more than I0 MPH above l i m i t ) F a i l i n g to respond to t r a f f i c Braking e r r a t i c a l l y Turning abruptly or i l l e g a l l y Driving into opposing or crossing t r a f f i c Driving with vehicle defect(s) AVERAGE C o e f f i c i e n t of c o r r e l a t i o n between Guide and test signals or signs

7
" 8

.70 .70 .65

65

.46 .36 .39

.79

9 10 11 12 13 14 15 16 17 18 19 2O 21 22 23

.65 .65
65

.62 .72
.69

.65 .65
55

.40 .67
.20

.55
55

.25
.67

Stopping i n a p p r o p r i a t e l y other than in lane

.55 .50 .50 .50

66

.47 .48 .62 .13

.56

.48 31

.51

were derived from data collected using procedures that were s u b s t a n t i a l l y d i f f e r e n t from those employed for c o l l e c t i n g data in the f i e l d t e s t . Guide p r o b a b i l i t i e s were based on data obtained by stopping each driver observed to be e x h i b i t i n g deviant d r i v i n g behavior and administering a breath t e s t to the d r i v e r . Observers accompanied patrol Field-test probabilities, officers for purposes of recording the data. on the other hand, were

obtained from data recorded on special forms during regular patrol by the patrol o f f i c e r s themselves. e a r l i e r study. Some detection procedures, such as the use of radar to detect fast speed, d i f f e r e d s u b s t a n t i a l l y from those used in the The three categories of BAC were estimated by the o f f i c e r however, verification of about twoeach time a d r i v e r was apprehended;

t h i r d s of the BAC ~ .10 estimates showed them to be 99 percent accurate. Second, Guide p r o b a b i l i t y tained d i r e c t l y from the Guide. values shown in Tables 6 and 7 were obThus, they were rounded-off values f o r

P(BAC ~ .10) that were extended to m u l t i p l e cue conditions and to P(BAC .05) through application of the simple adjustments presented at the bottom of the Guide. fact, Some loss in accuracy is l i k e l y to occur in the process. between f i e l d - t e s t values and the actual values In obcorrelations

tained from the e a r l i e r detection study were somewhat higher in a l l than those shown in Tables 6 and 7.

cases

During the f i e l d study 4019 apprehensions were made and information about each apprehension was recorded on a Drunk Driver Detection Guide and Record Form (see Figure 2). traced in Figure 5. The d i s p o s i t i o n of the 4019 apprehensions is Estimates were v e r i f i e d f o r As shown, 368 + 742 = 1110 drivers had an estimated

BAC ~ .05, and 742 had an estimated BAC ~ .10.

499 of the 742 by results obtained l a t e r from chemical t e s t s ; 492 of t h i s sample, 99 percent, were found to have a tested BAC ~ . i 0 . Detection data were combined from the e a r l i e r detection study (N = 643 apprehensions) and the field test (N = 4 0 1 9 apprehensions)

32

4019 APPREHENSIONS

2909 ESTIMATED BAC <.05

742 ESTIMATED BAC >__.10

368 ESTIMATED .05_<BAC <.10

I.
55 ARRESTED REFUSED TEST
499

ARRESTED BAC RECORDED

198 ARRESTED BAC NOT RECORDED

492 BAC_>.10

7 BAC <.10

Figure 5. Disposition of 4019 apprehensions recorded on Drunk Driver Detection Guide and Record Forms during three-month test period.

33

to cal,

provide

new cue d i s c r i m i n a b i l i t y and procedural

values

based upon a s u b s t a n t i a l l y thus provided by t h i s data in

larger data base (N = 4662 apprehensions). geographical, base, the r e s u l t i n g p r o b a b i l i t i e s Tables 8 and 9. Cues are l i s t e d p r o b a b i l i t y values.

Because of the larger numeri-

foundation

should be more r e l i a b l e and g e n e r a l i z These values are presented their in decreasing order for new DWI

able than those on the prototype Guide.

OPINIONS OF GUIDE UTILITY

Experienced police patrol after the Guide was introduced, officers were skeptical that use of the

Guide would enhance t h e i r DWI detection a b i l i t y . sample groups

Between four to six weeks of between two to seven A

p a t r o l o f f i c e r s within each agency were assembled to discuss t h e i r experience with the Guide and to obtain t h e i r opinions about Guide u t i l i t y . more systematic and comprehensive survey of p a r t i c i p a n t s was not authorized f o r the study. A content analysis of the recorded discussions is summarized in Table I0. Only two of the 10 groups had a consensus that the Guide would help them enhance DWI enforcement; four groups were s p l i t ment. tive Perhaps not c o i n c i d e n t a l l y , group consensus on t h i s point on t h i s issue; and four groups had a consensus that the Guide would not enhance DWI enforcenone of the four agencies with a negaincreased DWI arrest rate during the

t e s t period;

f i v e of the six agencies in which the group consensus was increased DWI arrest rate (see Table 1).

p o s i t i v e or s p l i t

In eight of the nine agencies in which there was discussion of the utility of the Guide for increasing patrol s e n s i t i v i t y to important cues, there was a p o s i t i v e or s p l i t consensus. t i v e opinions concerning the u t i l i t y aid in preparing DWI arrest reports, testimony. rated somehow in the Guide; There were also generally posiin providing were court split

of the Guide as a t r a i n i n g aid, as an and as an aid three

Five groups suggested that cue frequency should be incorpohowever, groups

34

TABLE 8 Cue D i s c r i m i n a b i l i t y Values Computed from 4662 Detections Made During Detection and Field Studies: P(BAC ~ . I 0 ) P(BAC ~ .10) WHEN CUE WAS ONE OF: VISUAL DETECTION CUES Cl C2 C3 C4 C5 C6 C7 C8 C9 C10 Cll C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 Turning with wide radius Straddling center or lane marker Appearing to be drunk Almost s t r i k i n g object or vehicle Weaving Driving on other than designated roadway Swerving Slow speed (more than i0 MPH below l i m i t ) Stopping (without cause) in t r a f f i c Following too closely Drifting Tires on center or lane marker Braking e r r a t i c a l l y Driving into opposing or crossing t r a f f i c Signalling inconsistent with d r i v i n g actions Slow response to t r a f f i c signals Stopping i n a p p r o p r i a t e l y (other than in lane) Turning abruptly or i l l e g a l l y Accelerating or d e c e l e r a t i n g r a p i d l y Headlights o f f (at night) F a i l i n g to respond to t r a f f i c Driving with vehicle defect(s) AVERAGE signals or signs Fast speed (more than 10 MPH above l i m i t ) lane ONE OR TWO OR THREEOR MORE CUES MORECUES, MORE CUES .63 .63 .60 .60 .58 .56 .53 .50 .49 .48 .48 .46 .46 .49 .42 .40 .37 .35 .32 .29 .18 .14 .07 .43 .63 .65 .66 .68 .62 .63 .55 .60 .73 .47 .51 .49
50
ii

.68 .67 .75 .70 .70 .69 .59 .71 .81 .52 .58 .52
.59

.54 .54 .63 .48 .48 .38 .45 .42

46

.65 .69 .65 .61 .51 .52 .62 .65

.60

.24 .54
I

.42 .63

35

TABLE 9 Cue D i s c r i m i n a b i l i t y Values Computed from 4662 Detections Made During Detection and Field Studies: P(BAC ~ .05) P(BAC VISUAL DETECTION CUES Cl C2 C3 C4 C5 C6 C7 C8 C9 C10 Cll C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 Turning with wide radius Straddling center or lane marker Appearing to be drunk Almost s t r i k i n g object or vehicle Weaving Driving on other than designated roadway Swerving Slow speed (more than 10 MPH below l i m i t ) Stopping (without cause) in t r a f f i c Following too closely Drifting Tires on center or lane marker Braking e r r a t i c a l l y Driving into opposing or crossing t r a f f i c S i g n a l l i n g inconsistent with d r i v i n g actions Slow response to t r a f f i c signals Stopping inappropriately (other than in lane) Turning abruptly or i l l e g a l l y Accelerating or decelerating r a p i d l y Headlights o f f (at night) F a i l i n g to respond to t r a f f i c Driving with vehicle defect(s) AVERAGE signals or signs Fast speed (more than 10 MPH above l i m i t ) lane 05) WHEN CUE WAS ONE OF:

ONE OR TWO OR THREE OR MORE CUES MORE CUES MORE CUES .8O .78 .76 .79 .77 .72 .69 .66 .61

.82 .79 .78 .79

.84 .80 .83 .79 .83 .81 .71

.83

77

.79 .71
77

.85 .65 .72

66

.72 .69 .74

.66

70

.71 .65 .69 .60

.75 .70 .70

68

.69 .77 .81

.70

57 48

.52 .50 .49 .42 .29 .23 .15 .59

.65 .64 .54 .60 .59 .66 .41 .70

.68 .69 .65 .76 .73 .74 .56 .74

36

TABLE I0 Summary of the Content Analysis of Group Discussions Conducted at Each Participating Agency Regarding Guide U t i l i t y (Group Consensus: ~ = Positive, ~)= Split, C ) = Negative)

AGENCY TOPIC Help enhance DWI enforcement Increases s e n s i t i v i t y to important cues Helpful as a t r a i n i n g aid Helpful i n r e p o r t i n g and t e s t i f y i n g Incorporate cue frequency in Guide A B C D E F G H I

0 ~ ~ ) 0 @ 0 0 ~ 0

@ 0 0 0 0~ (D @ @ @
~)

37

in opinion on t h i s suggestion.

The f o l l o w i n g s p e c i f i c changes in the Guide

were each recommended by persons in one o r two of the ten groups: Collapse weaving, d r i f t i n g , marker i n t o one cue. swerving and t i r e s on c e n t e r or lane

E l i m i n a t e p r o b a b i l i t y values on the Guide because t h e y might lead to d i f f i c u l t y in court t e s t i m o n y . Provide a method f o r mounting the Guide on a dashboard or v i s o r . Put Miranda and DWI warnings on the back of the Guide.

38

CONCLUSIONS AND RECOMMENDATIONS

The u t i l i t y resulted increase police traffic of the Drunk Driver Detection Guide f o r on-the-road This

detection of DWI was demonstrated in the f i e l d in an overall took place agencies in a field-test throughout

study.

Use of the Guide

increase in DWI a r r e s t rate of 12 percent. sample t h a t the United States, that

included 10 d i f f e r e n t employed

located

d i f f e r e n t types of p a t r o l s , conditions, DWI enforcement.

that included a wide range of geographic and levels of motivation f o r s i g n i f i c a n t changes

and that r e f l e c t e d d i f f e r e n t

Although there were no s t a t i s t i c a l l y

in detection practices,

such as those revealed by greater use of the more

d i s c r i m i n a t i n g cues or by arrests of more d r i v e r s with lower BAC levels, trends were in the expected d i r e c t i o n s . The DWI p r o b a b i l i t y values associated with the cues contained in the Guide were v e r i f i e d by f i e l d - t e s t results, providing a basis f o r using Guide Values with confidence. bility ity Although some m o d i f i c a t i o n s in Guide values Average p r o b a b i l -

were indicated, the overall r e s u l t was one of v e r i f y i n g the average probalevels as well as the values f o r i n d i v i d u a l cues. values calculated from f i e l d - s t u d y data were e s s e n t i a l l y the same as and Guide values f o r i n d i v i d u a l cues These r e s u l t s were obtained in spite of the

average Guide values, and f i e l d - t e s t were s i g n i f i c a n t l y c o r r e l a t e d .

d i f f e r e n t data c o l l e c t i o n methods employed in the o r i g i n a l detection study and in the f i e l d t e s t . Some d i f f i c u l t y little officers might be expected in gaining acceptance of the Guide Many feel detection they have is not a using i t , or t h a t

by police o f f i c e r s experienced in DWI enforcement. or nothing to learn from the Guide, primary problem in DWI enforcement.

On the other hand, a f t e r

stated that the Guide would be of value f o r reports , and providing

increasing patrol court testimony in

s e n s i t i v i t y to important DWI detection cues, t r a i n i n g inexperienced patrol o f f i c e r s , w r i t i n g DWI arrest conjunction with DWI arrests.

39

The Guide should be modified slightly, as shown in Figure 6. The DWI probability values in the modified Guide were based on data combined from the earlier detection study and the f i e l d test, providing a data base of 4662 detection events for these Values. The following three cues were
DWI and DWS:

eliminated because they did not discriminate much beyond chance between Fast speed (more than 10 M P H above limit) Failing to respond to t r a f f i c signals or signs m Driving with vehicle defect(s)

The modified Guide was further simplified by including only two instead of three adjustments: increasing values whentwo or more cues are observed, and estimating the probability of BAC equal to or greater than .05. Eliminating the adjustment for three or more observed cues should further f a c i l i t a t e the understanding and use of the Guide, and enhance the accuracy of adjusted values. Also, DWI probability values are stated as "chances in 100" rather than "percentages of" to avoid potential confusion between probabilities and expected frequencies. Modifications in the Guide should be reflected in the booklet designed to accompany the Guide. To support implementation of the Guide, a short, color, sound, 16-mm motion picture was produced. The film should be used along with the booklet to introduce potential users to the Guide in a cost-effective manner. The film summarizes how the Guide was developed, defines and illustrates the visual detection cues contained in the Guide, and describes how the cues should be employed for on-the-road detection of DWI.

40

DW] DETECTION GUIDE

Chances in 100 of nighttime driver with BAC equal or greater than .10

TURNING WITH WIDE RADIUS . . . . . . . . . . . . . . .

. ............

65 65

STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . . . . . . .

APPEARING TO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A L M O S T STRIKING ORJECT OR VEHICLE . . . . . . . . . . . . . . . . . . WEAVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60
60 60

DRIVING ON OTHER THAN DESIGNATED ROADWAY . . . . . . . . .

SWERVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SLOW SPEED (MORE THAN 10 MPH BELOW L I M I T ] . . . . . . . .
'4

55
55 50 50 50 50 45

STOPPING (WITHOUT CAUSE] IN TRAFFIC LANE . . . . . . . . . . FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DRIFTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TIRES ON CENTER OR LANE MARKER . . . . . . . . . ............

BRAKING ERRATICALLY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DRIVING INTO OPPOSING OR CROSSING TRAFFIC . . . . . . . . . . SLOW RESPONSE TO TRAFFIC SIGNALS . . . . . . . . . TURNING ABRUPTLY OR ILLEGALLY . . . . . . . " .........

45
45 40 35 30

SIGNALLING INCONSISTENT WITH DRIVING ACTIONS . . . . . 4 0 STOPPING INAPPROPRIATELY (OTHER THAN IN LANE] . . . . 35 ............... ACCELERATING OR DECELERATING RAPIDLY . . . . . . . . . . . . . .

HEADLIGHTS OFF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Special adjustment to the cue values 2 or more cues observed: add 10 to the larger value BAC equal to or greater than .05: add 15 to the value obtained for BAC equal to or greater than .10

30

Figure 6. Modified D r u n k Driver Detection Guide.

41

C~

APPENDIX
The Drunk Driver size, Detection Guide is illustrated inches). below in actual

8.73 cm x 12.38 cm (3-7/16 inches x 4-7/8 The p r i n t i n g

The Guide was

made of white p l a s t i c card stock. A booklet, "Drunk Driver

was in dark blue. of the Drunk The booklet and measured

Detection:

An Explanation

D r i v e r Detection Guide," was prepared to accompany each Guide. was p r i n t e d in blue on white paper, was stapled at the f o l d , 10.16 cm x 20.96 cm (4 inches x 8~ inches). pages. It

is shown on the f o l l o w i n g

DRUNK DRIVER DETECTIONGUIDE

Percentageof nighttimedriverswith BACequalto or greaterthan .10 Visual Cues STOPPING [WITHOUT CAUSE] IN TRAFFIC LANE . . . . . . FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . TURNING WITH WIDE RADIUS . . . . . . . . . . . . . . . . . . . . . APPEARINGTO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . DRIVING ON OTHERTHAN DESIGNATEDROADWAY . . . . STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . ALMOST STRIKING OBJECT OR VEHICLE . . . . . . . . . . . . SLOW RESPONSETO TRAFFIC SIGNALS . . . . . . . . . . . . . HEADLIGHTS OFF [AT NIGHT] . . . . . . . . . . . . . . . . . . . . . SIGNALLING INCONSISTENTWITH DRIVING ACTIONS . WEAVING "~ 70 60 60 60 55 55 55 50 50 45 45

TreES ON CEN'ER O, ILAN= : : : : : : : : : : : : : : : : : : : : :

DRIFTING SWERVING. . . . . . . . . . . . . . . . . ~. . . . . . . . . . . . . . . . . . . ACCELERATI't~(~"O'R'I)'E{;'E'I_'E'R'.~'I:I~I(3" i~APi[)I'Y ' i i i i i i i i i SLOW SPEED[MORE THAN 10 MPH BELOW LIMIT] . . . .
FAST SPEED [MORE THAN 10 MPH ABOVE LIMIT] . . . . . FALLING TO RESPONDTO TRAFFIC SIGNALS OR SIGNS . BRAKING ERRATICALLY . . . . . . . . . . . . . . . . . . . . . . . . . . STOPPING INAPPROPRIATELY[OTHER THAN IN LANE] . TURNING ABRUPTLY OR ILLEGALLY . . . . . . . . . . . . . . . . DRIVING INTO OPPOSING DR CROSSINGTRAFFIC . . . . . DRIVING WITH VEHICLE DEFECT[S] . . . . . . . . . . . . . . . .

4S
45 45 45
45 35 35 35 35 30 30 30

Me

SpecialAdjustmentsto the Percentages 2 cues: Add 5 to the largerpercentage. 3 or morecues: Add 10 to the largestpercentage. To predictBACequalto or greaterthan .05: Add 20 to the percentageobtainedfor driverswith BACequalto or greaterthan .I0.

A-1

W"

~J

INTRODUCTION
This booklet provides a detailed explanation of the visual cues contained in the Drunk Driver Detection Guide. These cues for discriminating nighttime drunk drivers from nighttime sober drivers were extracted from interviews with a wide variety of law enforcement specialists in drunk driver detection, from detailed analysis of over one thousand drunk driver arrest reports from different geographical regions, and from a field s t u d y in which cues observed in more than 600 patrol stops were correlated with driver BAC levels. Thus, the 23-cue Drunk Driver Detection Guide is the most systematically developed method currently available for visually predicting whether a vehicle operated at night is being driven by a drunk driver or a sober driver.

This booklet contains: A reproduction of the Drunk Driver Detection Guide A short explanation about the percentages presented in the Guide Explanations of the 23 visual cues used in the Guide

A-2

"

~PERCENTAGE INDICATORS
The percentage given after each cue in the Guide indicates the proportion of drivers on the average who exhibit that particular cue and who also have a Blood Alcohol Concentration (BAC) equal to or greater than .10. For example, the 70 percent following the first cue, Stopping [Without Cause] In Traffic Lane, means that out of 100 nighttime drivers who stop without cause in the traffic lane, on the average 70 will have a BAC equal to or greater than .10. Each percentage shown in the Guide is based on the observation of one cue. However, since more than one cue is often observed for a driver, the f o l l o w i n g simple adjustments are used to obtain percentages for multiple cues: If two cues are observed, find the larger of the two cue percentage values and add 5 to it. If three or more cues are observed, find the largest cue percentage value and add 10 to it. When you want to predict the proportion of drivers who have a BAC equal to or greater than .05, add 20 to the percentage value that was obtained in p r e d i c t i n g ~ . driver BAC equal to or greater t h a n .10. This applies to multiple cues as well as to single cues. Using the percentage indicators to decide whether or not to stop a particular driver will be a matter of department policy a n d / o r individual officer judgement. The Guide is only an aid that provides basic information concerning which visual cues are most likely to indicate a nighttime drunk driver.

DRUNK DRIVER DETECTION GUIDE

Percentageof nighttimedrivers with BAC equalto or greaterthan 10 - I Visual Cues STOPPING [WITHOUT CAUSE] IN TRAFFIC LANE . . . . . . FOLLOWING TOO CLOSELY . . . . . . . . . . . . . . . . . . . . . . . TURNING WITH WIDE RADIUS . . . . . . . . . . . . . . . . . . . . . APPEARING TO BE DRUNK . . . . . . . . . . . . . . . . . . . . . . . . DRIVING ON OTHER THAN DESIGNATED ROADWAY . . . . STRADDLING CENTER OR LANE MARKER . . . . . . . . . . . . ALMOST STRIKING OBJECT OR VEHICLE . . . . . . . . . . . . SLOW RESPONSE TO TRAFFIC SIGNALS . . . . . . . . . . . . . HEADLIGHTS OFF [AT NIGHT] . . . . . . . . . . . . . . . . . . . . . SIGNALLING INCONSISTENT WITH DRIVING ACTIONS .. WEAVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TIRES ON CENTER OR LANE MARKER . . . . . . . . . . . . . . . DRIFTING . . . . . . . . ~ ............................. SWERVING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ACCELERATING OR DECELERATING RAPIDLY . . . . . . . . . SLOW SPEED [MORE THAN 10 MPH BELOW LIMIT] . . . . FAST SPEED [MORE THAN 10 MPH ABOVE LIMIT] . . . . . FAILING TO RESPOND TO TRAFFIC SIGNALS OR SIGNS . BRAKING ERRATICALLY STOPPING INAPPROPRIATELY [OTHER THAN IN LANE] . TURNING ABRUPTLY OR ILLEGALLY . . . . . . . . . . . . . . . . DRIVING IMTO OPPOSING OR CROSSING TRAFFIC . . . . . DRIVING WITH VEHICLE DEFECT[S] . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . .

V
70 60 60 68 55 55 55 50 50 45 45 45 45 45 45 45 35 35 35 35 30 30 30

q,

Speczal Adjustments to the Percentages 2 cues: Add 5 to the

larger percentage. percentage ob-

3 or more cues: Acid 10 to the largest percentage To predict BAC equal to or greater than .05: Add 20 to the taineO tot drivers with BAC equal to or greater than .10.

A-3

tl

II 41.

=r

VISUAL CUE DEFINITIONS Stopping [Without Cause] in Traffic Lane

(Percentage: 70) 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. Intoxicated drivers might stop in lane when their capability to interpret information and make decisions becomes severely impaired. As a consequence, stopping (without cause) in the traffic lane is likely to occur at intersections or other decision points.

q.

Following Too Closely (Percentage: 60) The vehicle is observed following another vehicle ~vhile not maintaining the legal minimum separation. Turning With Wide Radius (Percentage: 60) During a turn, the radius defined by the distance between the turning vehicle and the center of the tui'n is greater than normal. This cue is illustrated below.

Appearing to be Drunk (Percentage: 60) 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: Tightly gripping the steering wheel Face close t o the windshield Eye fixation Slouching in the seat Gesturing erratically or obscenely Drinking in the vehicre Driver's head protruding from vehicle The drawing below illustrates the first three indicators in the above list.

. . . . .

#/

Driving on Other Than Designated Roadway (Percentage: 55) 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. The last example is illustrated on the next page.

A-4

I~

Driving on Other Than Designated Roadway

Headlights Off [At Night]

(Percentage: 50) The observed vehicle is being driven with both headlights off during a period of the day when the use of headlights is required.

Signalling Inconsistent With Driving Actions (Percentage: 45)

A number of possibilities exist for the driver's signalling 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, signalling opposite to the turn or lane change executed, signalling constantly with no accompanying driving action, and driving with four-way hazard flashers on. An example of this cue is illustrated b e l o w .

r_:

.1,

Straddling Center or Lane Marker

(Percentage: 55) The vehiole is moving straight ahead with the center or lane marker between the left-hand and right-hand wheels.

,'. ,~',.'1:'.'"~ ..........

, .......

"',":":':'.!!I!

!i'~'""' "':"! ''''~''''~ ....... ~' "i

Almost Striking Object or Vehicle

(Percentage: 55) 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.

Slow Response to Traffic Signals

(Percentage: 50) The observed vehicle exhibits a longer than normal response to a change in traffic signal. For example, the driver remains stopped at the intersection f o r an abnormally long period of time after the traffic signal has turned green.

Weaving (Percentage: 45)

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

-4.

A-5

another. Weaving drawing below.

is illustrated

by the

Drifting ,i,,.,,,i:~,&.,i'~;~':.:.,;..,.:,,::....'.:,.":.: i" :d'"............

/
g',

S w e r v i n g (Percentage: 45)

Tires on Center or Lane M a r k e r

(Percentage: 45) The left-hand set of tires of the observed vehicle is consistently on the center line, or either set of tires is consistently on the lane marker.
D r i f t i n g (Percentage: 45)

I
L

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 on the next page, 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.

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 o f f 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 o p p o s i n g traffic.

A-6

~J

i.

Accelerating or Decelerating Rapidly

(Percentage: 45) 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.

prohibited zone, at a crosswalk, far short of an intersection, on a walkway, across lanes, for a green traffic signal, or for a flashing yellow traffic signal. The drawing below shows one example of this cue.

vI,;

Slow Speed [More than 10 MPH Below Limit] (Percentage: 45)

The observed vehicle is being driven at a speed that is more than 10 MPH below the speed limit.
4

Fast Speed [More than 10 MPH Above Limit] (Percentage: 35) The observed vehicle is being driven at a speed that is more than 10 MPH above the speed limit. Failing to Respond to Traffic Signals or Signs (Percentage: 35) The observed vehicle falls to respond to a traffic signal or sign. For example, the vehicle fails to stop for a red traffic signal, fails to stop for a stop sign, or fails to slow for caution signals. Braking Erratically (Percentage: 35) The driver of the observed vehicle is braking unnecessarily frequently, maintaining pressure on the brake pedal ("riding the brakes"), or braking in an uneven or jerky manner.

Turning Abruptly or Illegally

(Percentage: 30) The driver executes any turn that is abnormally abrupt or illegal. Specific examples include turning: with excessive speed, sharply from the wrong lane, a U illegally, and outside the designated turn lane. This cue is illustrated below.
I! r!ll!mi= ILl ! I'lL! I ~ I'' L" ' ' ~ql I" " !' '
'""' ~

Stopping Inappropriately [Other Than in Traffic Lane] (Percentage: 35)

The observed vehicle stops at an inappropriate location or under inappropriate conditions, other than in the traffic lane. Examples include stopping: in a

A- 7

gk,,

a~

Driving Into Opposing or Crossing Traffic

"~ ~, (Percentage: 30) The vehicle is observed heading into opposing or crossing traffic under one or more of the following circumstances: driving in the opposing lane, driving the wrong w a y on a one-way street, backing into traffic, failing to yield the right-ofway. The last circumstance is illustrated below.
_ ~ ~ , . ,.,~,1 ,' ~ i ' F

.~...~

,.

,~

,//

../x
Driving with Vehicle Defect[s]

JE

(Percentage: 30) The observed vehicle is being driven with one or more defects, such as: faulty headlight, faulty taillight, flat tire, or one of many other observable mechanical or electrical defects.

4,.

A--8