HUMAN FACTORS REQUIREMENTS FOR REAL-TIME MOTORIST INFORMATION DISPLAYS
"COWRAD [. DUDEK
VOL, 9 A STUDY OF PHYSICAL DESIGN REQUIREMENTS FOR MOTORIST INFORMATION MATRIX SIGNS
. C. J. Messer · W·.· R •. Stockton .
J. M. Mounce Contributions by
D. A. Andersen @] J. M. Turner
[f@[p©[fU . · from·theTexasA&M
·RESEARCH FOUNDATION College Station, Texas
· Texas Transportation Institute Texas A&M University
. College Sta ti on, Texas 71843
Prepared for
I
Federal Highway Administration Offices of Research and Development
Contract No. DOT-11-8505
· .· F_ebruary 1978 ·
Technical Report Documentation ?:ige
l. Repo't No. 3. "eco.p1"ent's Catalog No. ~----1
~~·oe-nrn-<~A-c-ce_'_""":_o_. -----1-- ~ ______ J FHHA-RD-78-lj ---- 5. Repo't Date
HUMAN FACTORS REQUIREMENTS FOR REAL-TIME MOTORIST I February 1978 -INFORMATION DISPL Vol. 9 A Study o
AYS I 6. Performing Orgoni zotion Code
Motorist f Physical Design Requirements for
~ Pedo,ming-O,gani zo!;o:i Report No. Information Matrix Signs 1 >-, • .:,,. •' C. J. Messe r, w. R. Stockton, J. M. Mounce
I ca~tributions by D. A. Andersen, J. M. Turner r--------------- ·-----------------~--
, ., and Adcires.s
1
. 9. ?er,•crm;ng Otgon;2at1or\ Nor.i
Texas Transportat Texas A&M Univers Co 11 ege Sta ti on,
ion Institute ity Texas 77843
with
10. Work Unit No. (TRAISJ
11. Contract or Grant No.
DOT-FH-11-8505 13. Type of Report ond Period Covered
12. Sponsoring Agency Name on d Address
Department of Transportation Federal Highway Administration
, Final Report j (June 1974-February 1978)
I Offices of Research and Development
rWashington, D.C. 20590 Suppl .. oentory Notes
FHWA Contract Manager: Truman M. Mast (HRS-31) Principal Investigator: Conrad L. Dudek
p Sponsoring Age,,cy Coae
I
Co-Principal Investigator: R. Dale Huchingson 16
· Abstroc1 The purpose of this study was to identify the physical design requirements o-fi
(hangeable message lamp matrix signs. Design characteristics addressed incluce sequential and run-on displays, message length, words per line, number of li~es, display rates, legibility, bulb loss, and symbolic substitution. Effects of driver work load were also addressed.
The studies were conducted first in a laboratory equipped ~·lith a rear orojection
lscreen and subject resoonse controls. Slides and motion picture fi1r were used with controlled exposures to obtain subject responses. Percent correct ~esoJnse was the
levalu~ti?n criterion for all variables except symbolic substitution which related to assoc1at1on. I Selected laboratory results 1t1ere subjected to limited field valid:i.tion in a cor-1trolled drivir.a environment. The field studies considered subject oerfornance under 'both "unloaded;; and "loaded" conditions.
Sequential display formats appeared to be considerably superior to rnn-on formats Several combinations of message length, words per line and number of lines weret2sted Certain combinations showed marked superiority over other combinations. A disolay rate of 0.5 seconds per vJOrd appeared to be the fastest acceptable for four-word messages, with 1.0 seconds/l·rnrd very near o;=:ti':ium for both four-and eig!-;:-,.,:ord i:'.25-
lsages. The 85th percentile legibility distance of the 18-inch lamp matrix sign \·1as iabout 630 feet, or 35 feet/inch of letter height. Effects of bulb loss \·;ere more pro!nounced on unfamiliar drivers than on familiar; 10 percent loss appeared tolerable. !Acceptable association vvas found betv1een matrix symbols and graphic symbols, except for r"rved ~bol s I ,,_. \....~ ,, ,.-;-::-----
18. Distribution STotem~nt • 17. iC ~y Word,
i Dynami~ signs_, matrix signs, var~able No restrictions. This document is !message signs, changeable message s1gns, available to the public through the ,dynamic displays, display rates, legi- Nati ona 1 Technical In format i on Service, !bil~ty distance, bulb loss, message Springfield, Virginia 22151 ilenath-, i ~
syrnob l i c substitution ~ 9. :~-c-urdy Clo~sd. ~of this 1epo,t) 20. Security Classif. (of th;s pog~) 121. No. of Pages 22. Price
' I J::cl ass ifi ed Unclassified i I
Form DOT F 1700.7 \8-72) Reproduction of complated page authorized
PREFACE
This document is part of a seventeen-volume report entitled, Human
[?.c!:9I~Reguirements For Real-Time Motorist Information Displays. Titles of
al 1 volumes are shown below.
FHWA-RD Volume Nurr.ber ----
1 78-5
2 78-6
3 78-7
4 78-8
5 78-9
6 78-10
7 78-11
8 78-12
9 78-13
10 78-14
11 78-15
12 78-16
13 78-17
14 78-18
15 78-19
16 78-20
17 78-21
Title
Design Guide
State of the Art: Messages and Displays in Freeway Corridors
Summary of Systems in the United States
Bibliography and Selected Annotations: Visual Systems
Bibliography and Selected Annotations: Audio Systems
Questionnaire Survey of Motorist Route Selection Criteria
Analysis of Driver Requirements for Intercity Trips
Analysis of Driver Requirements for Intracity Trips
A Study of Physical Design Requirements for Motorist Information Matrix Signs
Human Factors Evaluation of Traffic State Descriptor Variables
Human Factors Evaluation of Route Diversion and Guidance VariablE:s
Supplement to Traffic State Descriptors and Route Diversion and Guidance Studies
Human Factors Evaluation of Audio and Mixed Modai Variabies
Point Diversion for Special Events Field Studies
Freeway Incident Management Field Studies
Feasibility of Audio Signing Techniques
Driver Response to Diversionary Information
ACKNOWLEDGMENTS
Special acknowledgment is made to Dr. Truman Mast, FHWA contract Manager,
for his invaluable counsel, advice~ and guidance throughout this project.
Dr. Mast worked closely with the research staff and was always available to
share ideas and offer constructive critique, adding more depth and dimension
to this research project. His associates, particularly Jim Ballas and Joe
Peters, are recognized for their technical consultation and constructive
criticism. Dr. John Eicher provided FHWA administrative support on this
project. Acknowledgment is also made to Lawrence D. Powers, Lyle Saxton, and
Samuel Tignor who, together with Truman Mast and John Eicher, developed the
real-time motorist information display research program.
The authors also wish to express their appreciation to Dr. Conrad L.
Dudek, the Project Principal Investigator, and Dr. R. Dale Huchingson, Co
Principal Investigator, for their advice and guidance, and to Mr. Donald R.
Hatcher for coordinating the preparation of the final document.
B:ickground
Problem ..
Objectives
Research Approach.
Introduction
Media-Master Laboratory.
Test Subjects ....
TABLE OF CONTENTS
Laboratory Evaluation of Matrix Displays
Experimental Factors
Experimental Design.
Experimental Development
Experimental Administration.
Data Reduction
Data Analysis.
Sequential Sign Results.
Run-On Sign Results ...
Summary of Laboratory Results of Matrix Displays
Field Evaluation of Matrix Signs
Introduction
Purpose.
Scope.
:~1atri x Sign.
:::xnrirnental Design.
1
2
4
6
6
7
8
12
14
18
20
21
21
22
33
40
40
40
41
41
43
TABLE OF CONTENTS (continued)
Experimental Administration
Data Reduction.
Data Analysis .
Laboratory Evaluation of L2mp Matrix Bulb Loss
Introduction ....
Experimental Development.
Experimental Administration
Data Reduction and Analysis
Results . .
Conclusions and Recommendations
Laboratory Evaluation of Symbolic Matrix Substitution.
Introduction ....
Experimental Design
Experimental Development.
Experime~t2l Administration
Data Reduction.
Data Analysis .
Summary of Results.
Study Summary ..
Conclusions
Design Recor::rcerda ti ons.
References . . . . . . . . .
v
44
46
47
57
57
57
61
62
68
71 I .l
73
75
75
82
82
84
91
94
99
103
TABLE OF cmnE::IS (continued)
A~:~~dix A - Experimental Desig~ for Laboratory Evaluation
:.= ··:.:.trix Displays ...
A~pe~dix 8 - Experisental Design for Field Evaluation of
~atrix Signs ..... . -~
.L.ppendix C - Experirrienta1 Cesign for L2'::oratory E'1aL2tion
of Lamp Matrix Bulb Loss ...
Appendix D - Experinental Design of La~oratory ~v21j2tian
of Sy::~bo1ic ;•a:rix SubstitJtion.
~atrix and Graphic Symbols
104
133
1 ,,.., J.,"...:
LIST OF TABLES
Biogra phi cal Data of Laboratory Test Subjects
Table 2. Percent of Drivers 18 Years of Age and Older
Completing Education Level Shown .....
Table 3. Biographical Data of Field Test Subjects ..
Table 4. The Effects of Number of Lines Versus Number of
Exposures Over a Range of Display Rates as
Measured by Relative Performance Ranking ..
Table 5. ·Summary of Results of Field Studies for Electronic
Matrix Signs. . . .
Table 6. Rank Order of Four-~·Jord Sign Designs at 0.5 Seconds
Per \'lord Rate
Table 7. i·Jord Lists Used in Bulb Loss Study.
Table 8. Summary of Percentage Bulb Loss Versus Percentage
Correct Response Ve rs us tford Length . . . . . .
Table 9. Bulb Loss Percentage Associated with Criterion
Performances of 853 and 95% as a Function of
Table
Table
\ford Length . .
Selected Human Factors Requirements Design
Criteria for Matrix Sign ..... .
Recommended Minimum Letter Heights for Various
Message Lengths, Sign Locations and Approach Speeds
for Familiar and Unfamiliar Drivers ....... .
vii
. .
Page
9
10
11
28
43
52
59
63
69
100
101
\,
LIST OF FIGURES
Fig~re l. Four-Word Sequential Signs Displaying First
Exposure .
Figure 2. Eight-Word Sequential Signs Displaying First
Exposure .
Figure 3. Best Four-Word Message Sign Display Beginning
at First of Message ..
Figure 4. Performance of Same Sign Displaying Four-Word
and Eight-Word Messages From First of Message.
Figure 5. Effectiveness of Two-Line, One \·lord Per Line Sign
and One-Line, TvJO \ford Per Line Sign ....
Figure 6. Effect of Starting Position and Display Time of
Message Upon Performance (Two and Four Exposures).
Figure 7. Effect of Starting Position and Display Time of
Message Upon Perforriance (One Exposure).
Figure 8. Effectiveness of Six and Ten Character Run-On
Displays Starting at First of ~1essage.
Figure 9. Effectiveness of Run-On Signs for Four and Eight
~lord Messages Starting in the Middle . .
Figure 10. Effectiveness of Displays for Four and Eight
Word Messages Starting at First of Message .
Figure 11. New Trailer Mounted Lamp Matrix Sign Used in Field
Studies ....
t1aure 12. Vehicle Within Lane of Cones 11 Loaded 11 Study.
Figure 13. Comparison of Field and Laboratory Results for a
Selected Sign Display ..
.
.
. . . .
16
17
23
25 '
27
31
32
34
36
38
42
45
50
LIST OF FIGURES (continued)
Figure 14. Comparison of Visual Acuities of Subjects
with National Population ....
Figure 15. Legibility Distance Observations
Figure 16. Bulb Loss Versus Percent Correct Response as a
Function of ~lord Length in a "Familiar" Motorist
Condition. . . . . . . . . . .
Figure 17. Bulb Loss Versus Percent Correct Response as a
Function of ~·Jard Length in an "Unfami 1 i ar"
Motorist Condition . . . . . . . .
Figure 18. Bulb Loss Versus Percent Correct Response as a
Function of \~ord Length in an ''Average''
Motorist Condition ....
Figure 19. Highly Recognizable Symbols.
Figure 20. Group I Symbols - Smoothed Curved Line Forms
Figure 21. Group I I Symbols - Outline Forms . .
Figure 22. Group II I Symbo 1 s - Line or Sno1·1fl ake Forms.
Figure 23. Group IV Symbols - Other Forms . . . . .
Figure 24. Group IV Symbo 1 s - (Cont'd.) - Color Substitutions
Figure 25. Examol e Symbols.
Figure 26. Group I Syr~bo 1 Selection Results
Figure 27. Group II Symbol Selection Results.
Figure 28. Group I I I Symbol Selection Results
Figure 29. Group IV Symbol Selection Results.
Figure 30. Group IV Symbols (Cont'd.) - Selection Results
. . .
. . .
54
55
,.. 11. 0 .
65
67
74
77
..,0 /u
79
80
83
85
86
87
89
90
INTRODUCTION
Backc;round __ ....::::::...._. ___ ___
~s traffic demand increases on the freeways, the possibility of freeway
ccr";2stion greatly increases. In the last ten years the technique of meter-
ing freeway demand has been used effectively to improve the operations on
some crowded freeways under normal operations. However, when an accident
occurs on the freeway, the demand exceeds the freeway capacity at the
11 bottleneck" area, and major freeway congestion occurs.
The introduction of real-time motorist information systems has provided
traffic engineers the opportunity to redistribute traffic around an incident,
thereby reducing travel time and improving the level of service on the free-
way. The use of changeable message signing systems has been proposed as an
effective means of displaying real-time motorist information along high volume
freeways. The Texas Transportation Institute in cooperation ~vith the Federal
Highway Administration is researching the human factors requiremen~s for using
changeable message signs and other motorist information systems on freeways.
This report describes one part of the research. Future research dealing with
audio signing will be reported in a later volume of the reports emanating
from this project.
Electronic, changeable message signing systems offer numerous and ~ro~ising
alternatives to the traffic engineer in providing the freeway motorist with
timely traffic infomation. These dynamic information displays appear to pos-
sess the typ~ c1 F 1~'essage display features necessary to provide the variety of
infor:iation needed by motorists traveling on urban freeways. Recognizing the
apparent features and advantages of electronically-operated, motorist informa-
tion 'oJStems, several types of displays have been developed by commercial firms
1
and operated by traffic engineers. These systems have included a wide
range of display forms such as lamp matrix, rotating drum, magnetic disc,
and blankout signs. Volume I of the reports emanating from this project
pr2sents a full description of the various displays used throughout the
country (l).
The research described in this report includes only lamp matrix
changeab 1 e message signs. There 1t1ere several reasons for this decision.
A changeable message, matrix sign was available to the study and could
be used at little cost. The sign could display both sequential (flashing)
and run-on (moving) messages. In addition, matrix signs presently con-
stitute a large segment of motorist information and co~rnercial outdoor
advertising signing and can be expected to continue to play a leading
role in the foreseeable future. Other types of display systems have
been utilized successfully, and new display methods undoubtedly wi11 be
developed in the future.
Problem
Since the use of changeable message matrix signs as dynamic motorist
information displays is a relatively new concept dealing with a wide range
of complex issues, very fe\'i design and operational guidelines are avai1ab1e
to insure effective utilization of their capabilities. Available static
sign design criteria on legibility and word reading speed do not necessari1y
apply to dynamic matrix displays. Even less is known about the interrela-
tionships that may exist among important sign design variables, such as
message length, layout, and display technique, and the human factors
characteristics of message reading speeds and information retention
• ~ 1 ~ .J.. ~ capat11.11..1es.
2
The design of a matrix sign for use in a freeway motorist information
system requires that the engineer knmv: 1) the information to be transmitted
by the sign, 2) the driver 1 s ability to read and comprehend it within a given
t·; ::~ r-rame, and 3) the cost of providing the information via that particular
s·iqr1 design. Having this information, the engineer could select the overall
sign system design that would conceptually
a) Maximize the information transferred for a given cost of the
signing system.
b) Minimize the cost of the signing system for a given info~mation
transfer level required.
c) Evaluate other trade-off options between information transfer
and system costs.
It was known that the cost of matrix signing systems varied with the size,
message configuration, method of operation and number of signs used. While the
exact cost of various design alternatives might not be known~ priori, the
engineer could determine them by requesting alternative bids on a proposed
contract. Other activities within this overall research project seek to deter
mine the infoi'i~;ation that v10uld need to be transmitted to freeway rr:otorists
under a given set of circumstances. Thus, this particular report will focus
on evaluating the drivers• ability to read and, in general, comprehend various
traffic messages within a given time frame as a function of the type of r:;atrix
sign display used.
In the routine daily operational environment of motorist information sign
ing systems, it is reasonable to expect that electronic matrix signs must func
tion while experiencing a small percentage of bulb loss (outage) at any
or-e ':'~r-:e. It was desired to determine the reduction in word readability that
3
may occur due to a reduction in the percentage of bulbs operating caused by
random bulb failures. Acceptance criteria for defining when bulbs should be
replaced, based on percent bulb loss, are also needed to assist operating
a;encies in establishing efficient bulb replacement programs.
The use of electronic matrix signs to encourage diversion of motorists from
a pr·imary route to an alternate route is a frequently discussed control option.
As it is likely that some of the diverted drivers would be unfamiliar with the
alternate route, it may be necessary to provide guidance information along the
route. This guidance would likely take the form of trailblazers: One approach
is to trailblaze the route with a symbol. Using this technique, the electronic
matrix sign would display !I ••• FOLLOW (s11mboZ). 11 Symbolic trailblazers would
then be placed along the alternate route to guide the diverted motorist to
his destination or back to his primary route. However, symbols formed on a
matrix sign are not exact replicas of the same symbol formed on a painted sign.
There is a need to evaluate the ability of drivers to correctly associate
symbols formed on a matrix sign with corresponding symbols formed on a painted
static sign (trailblazer).
Objectives
The overall goal of this research effort was to develop technical data
related to human factors physical design requirements of motorist informa
tion matrix signs. To meet this goal, the following objectives were
established:
l. Determine how sequentially exposed messages compare with continuously
moving or 11 run-on 11 messages in terms of subjects' ability to repro
duce the messages that were presented for the same length of tine.
4
2. Evaluate the effects that message word length has on subjects•
ability to reproduce the message. Consider four- and eight-word
messa·;es.
3. Determine how message display time and resulting word display rates
(for a given message length) affect subjects' ability to reproduce
words. Evaluate word display rates of approximately 1.0, 0.5 and
0.25 seconds per word.
4. If messages of a given word length are exposed sequentially within
a given time frame, determine the display 1 s effectiveness when one
word per line and two words per line are used as reflected by subject 1 s
ability to reproduce the messages.
5. If messages of a given word length are exposed sequentially within a
given time frame, determine the effectiveness of one-, two- and four
line displays.
6. Determine how starting a display in the midd.le of a message affects
performance as compared with starting at the beginning of the mes
sage sequence.
7. Determine the effects that various percentages of bulb loss (la~p
outages) have on word readability.
8. Evaluate the ability of subjects to correctly associate symbols
formed on a matrix sign vii th corresponding sy:::Dcl s formed on e.
painted static sign.
5
RESEARCH APPROACH
Introduction
The research approach selected for evaluating the desired human factors
performance characteristics of electronic matrix signs consisted primarily of
·laboratory studies followed by selected controlled field testing using a new,
fully operational, full-scale matrix sign. Four studies were conducted:
1. Laboratory Evaluation of Matrix Displays - Human factors performance
characteristics of matrix signs measured in laboratory.
2. Field Evaluation of Matrix Signs - Selected laboratory tests checked
for validity in full-scale field study.
3. Laboratory Evaluation oflamp Matrix Bulb Loss - Readability of signs
having given percentages of lamp outages.
4. Laboratory Evaluation of Symbolic Matrix Substitution - Association
between matrix and painted sign displays evaluated.
All laboratory testing included visual simulations of electronic .._ . na 1..ri x
signing displays. In most cases, 35-mm slides or 16-mm motion picture film of
a full-scale, trailer-mounted matrix sign were used to increase the fidelity
and realism of the laboratory studies. Movies of various matrix signing dis-
plays were used exclusively for the evaluation of matrix displays; whereas,
slides of signs were used in the other laboratory studies. Some artwcrK was
used to simulate static sign displays in the symbolic substitution study.
A trailer-mounted lamp matrix sign obtained from an electronics firm
located in Texas \'las used in the laboratory studies. The sign was composed
of a 7 x 60 array of 25-watt bulbs, 1.5 feet high by 12 feet long. Any message
or symbol not exceeding about 10 characters on a single line could be displayed.
6
Normally, a character was five bulbs wide. Programming of the sign was accom-
p1ishE:d by punched paper tape. Characters 1•1ere formed by one vertical column
of bu.lbs at a time, i.e., each column of holes on the tape correspond to a
cJ1u~n on the sign. The punched tape, therefore, is a replica of the charac-
t.2r:; that 21re displayed on the sign.
The operational features of the sign allowed messages to be presented in
two basic forms: 1) sequential and 2) run-on. Sequential displays present a
message by dividing the message into sections and exposing each section in a
sequence of discrete displays. For examp 1 e, a four l'iOrd message ·di sp 1 ayed on
a one line sign could be presented by displaying the first t\-10 1,11ords fol1owed
by a display of the last two words of the message. Run-on, or moving messages
present a message by moving the message continuously across the sigh display
from right to left until the message has been completely displayed. The speed
with which messages were displayed was regulated by a calibrated dial-potentio-
meter which was an integral part of the matrix sign control system. Since the
sign had only a one-line display, photographic means were employed to si~ulate
up to four lines of display. Field performance testing of a new computer pro-
grarnmable, twu-~ine n~trix sign was ccnducted at the Texas A&M Research Annex.
Media-Master Laboratory
It l'ias recogn·ized that the laboratory studies nee·::ed to be as rea1-wor1d
as possible. However, experimentation with a 1arge number of human subjects
also required expediation. ''. The media-master laboratory located on the Texas \
A&M University campus provided an excellent facility to conduct human experi-
ments of this nature. The laboratory provides remotely-controlled environmental,
te.:;tiqg and e·1ahiation capabilities for approximately 20 subjects, if necessary.
7
Film and slide presentations used in this study were projected onto an
opaque wall screen using the rear-projection method. Taped voice instructions
and the silent black and white film were synchronized together by an Edex
multi-channel control system located in the rear-projection room.
All of the 226 subjects tested in the laboratory studies were selected
from residents of Bryan/College Station, Texas. The biographical characteristics
of these subjects were stratified as to age, sex, education, and mileage
driven per year as shown in Table l. The characteristics of the population
pool were formulated carefully to be representative of the national driving
public (Table 2).
Subjects for the field study were selected from a pool of TTI personnel
not associated with the project which was stratified according to age, sex,
and education as shown in Table 3. ;These subjects were chosen to rep"licate
as closely as possible, within the confines of the subject pool, the distri
bution of the national driving public. As will be described in detail later,
the field study was of limited size and limited objectives, primarily to
validate the ranking and scale of percent correct response of selected labor
atory studies. As a consequence, only 20 test subjects \'/ere used.
8
l..O
f\gc·
Sc:x
Educational f.<•ve l
Miles Driven Per Year
Table 1. Biographical data of laboratory test subjects
18-'.3·1 2:)<J1 35-tl4 ,15-54 55-64 Over G4
11.7% 21. 11% 19.8% 27.9% 13.5% 2. 7();,.
~la le Female
70. 39{, 29.7%
,_
Elcm(;ntary Junior High
] ?, " 4 5 6 7 8 9 ,)
0 . D':) 0 . 9~{) 3.G% 3.6% 1.5% 7.2% 11.7%
High School College
10 11 1 ., L, ]_ 2 3 4
11.7% 12.n% 37.8% .9% 2.7% 1. 8% 0
0 - 10,000 10,000 - 20,000 over 20,000
·1 2. ()f.'{) 42.G% 14.8% --·-... ·~·--.--.1---·--··-~- ·-~··-... ...., ______________________ ,.
TABLE 2 PERCENT OF DRIVERS 18 YEARS OF AGE AND OLDER
COMPLETING EDUCATION LEVEL SHOWN*
MALES Ag·e Groues Elementary H1qh School Coll eae
1-3 4 1-3 4 or more
lB-24 3 6 4 2 1
25-34 1 2 4 2 l 2
35-44 l 2 3 l I 1
45-54 2 2 3 1 l
55-64 2 1 2 1 1
over 64 4 l 1 1 0
Total Males 13 14 17 8 6
FEMALES
18-24 2 4 3 2 I 0
25-34 1 1 3 1 l
35-44 l 1 3 l l
45-54 2 l 2 l 1
55-64 2 1 1 l 0
over 64 3 l 1 0 0
Total Females 11 9 13 I 6 3
GRAND TOTAL 24 23 30 14 ! 9 l '
Totals
16
11
8
9
7
7
i 58
11
7
7 7 ,
I 5
5
42
100
*Adopted from United States Statistical Abstract, U. S. Bureau of the Census, Washington. D. C., U. S. Printing Office, 1971, and Highway Statistics, U.S. Department of Transportation, Washington, 0. C., U. S. Printing Office, 1973.
10
........
........
Age
Sex
Educational Level
T0-blC' 3. Biogro_phicul du.tu of field test subjects
--·------.-· . ---- __ ........ __
18-24 25-34 35-44 45-54 55-64
25% 50% 15% 5% 5%
Mule Female
55~{, 45'.i;
High School College
35;:. 1--3 years 4 or more
50~~ 15%
LABORATORY EVALUATION OF MATRIX DISPLAYS
Ex~erimental Variables
As noted in the list of objectives of this study, several different,
bJt ~nterrelated, variables were to be evaluated. A brief discussion of
these factors will be presented subsequently to aid in the understanding
of their meaning and to identify why a given range of values of each was
selected for study in the media-master laboratory.
Display Types - As noted previously, two types of message displays
were tested: (l) discrete or sequential message displays and (2) run-on
or horizontally moving message displays. Almost all electronic matrix
signs use one of these two types of message display techniques.
Starting Point - An important, but controversial, issue arose during
the formulation of the experimental design with regard to simulating and
evaluating 11 average'' sign reading time conditions. Reading rates could be
closely controlled and measured by starting all messag~ displays at the
first of each message and showing a first-to-end-of-message display sequence.
However, it is possible that an urban freeway motorist might start reading
the sign anywhere during the message display sequence. If this occurred,
then the driver would have to re-read most, if not all, of the subsequent
message sequence to understand the message. To study this situation, ha1f
of all messages were started at the first while the other half began at
the middle of the message (the remainder of which vrns then displayed) and
then the entire message was repeated.
Message Length - In order to keep the factorial experiment to some
r:12r;ageable size, on1y two message lengths 1t1ere selected. These were
12
four-\,1ord and eight-word messages. To il 1 ustrate, a four-word message was:
11 SHARP TURN NEXT RIDGE u
and an eight-word message was:
11 SHARP TURN NEXT RIDGE H\1Y-5 E.AST SLOW DOvJW
In general, four-word messages contained information on a 11 situation 11
and its 11 1 ocati on 11 v;1hi le eight-word messages contained this information
plus additional information on a given "traffic audience" and the desired
"action" to be taken. All of these four message "units" were described
by two \vords, normally four or five 1 etters long, except for the "traffic
audience" where numbers were used in half of the "unit," e.g., HHY-5 EAST,
US-69 WEST. Word lengths of four or five characters were chosen so that two
similar words could be displayed simultaneously on the same line of the ~atrix
sign. Within these constraints, four-word and eight-word messages were selected
to be equi-difficult in readability and recall.
Words Per Line - In the anticipated freeway traffic information usage in
which the signs were being tested and evaluated, most traffic messages that
might be formulated in practice can be conveniently structured in either one
or two word "units 11 requiring 10 to 12 characters. This \·las fortunate
since the matrix sign used in the laboratory studies was limited to ac>Jt
l 0 characters. Thus, it \vas decided to study both one and b-10 \vord per 1 ine
displays for both sequential and run-on signs. A one-word, run-on sign ~as
assumed to have a six-character display while a tvo-word, run-on sign had
ten characters.
Number of Lines - One major design variable for both sequential and run
on matrix sign dispiays is the number of lines of message display. ~/hi1e one
or two-line signs aopeared to be more frequently used in the United States
13
today, some four-line signs are being used (1J. It was therefore decided to
test one-, two- and four-line sign displays for sequential messages. Only
one-line "run-on" displays were tested .
.Q_isplay Rate - The slowest speed at which messages being displayed
could be read was defined as the ~isplay rat~1
of the message display in
seconds per word (seconds/wo~d). A four-word message displayed one time
from start-to-end in 2.0 seconds would have a display rate of 0.50 seconds/
word (2.0 seconds/4 words). The literature indicated that word reading
rates of 0.3 to 0.5 seconds per word could be expected for words having
one or two syllables when used in the sequential mode (~, ~). On the
other hand, some local TV stations had found that moving messages on TV
appeared more acceptable when displayed at 6 characters per second rather
than 12 characters per second. Since a blank space is assumed a character,
these rates would correspond to approximately 1.2 words/sec. (0.83 seconds/
word) being acceptable and 2.4 wordsf~ec. (0.42 seconds/word) probably not
being acceptable for run-on or moving message displays using the laboratory
matrix sign. This assumes that the 10-character matrix sign would produce
one word for every five characters. In summary, it was decided to test
basically three word rates of one, two and four words per second for both
sequential and run-on message displays. Equivalent display rates are 1.0,
0.5 and 0.25 seconds per word.
Experimental Design
The experimental design of this study of 11 Message Displays 11 could be
considered a 11 rncdified 11 randomized factorial design. The factors and their
l eve 1 s of treatment that were tested \<Jere:
14
Type - (2); sequential, run-on
Start - (2); first or middle point of message
Length - (2); four or eight words
Words - (2); one or two per line
Lines - (l); one-line (run-on)
Lines - (3); one, two or four-line (sequential)
Exposures - (3); one, two or four exposures (sequential)
Rate - (3); 0.25, 0.5 and 1 .0 seconds/word
All of these studies were replicated once using a different message for each
sign to try to reduce the effect that a particular message might have on a
sign test condition.
There were five, four-word sequential sign designs tested as shmm in
Figure 1. Three sign designs had one word per line and two signs had two
words per line. In addition, one, two and four line displays were tested.
The first exposure of the four-vrnrd message, "SHARP TURN NEXT RIOGE 1' is
being displayed on each sign. Signs S-U.-4-1 and S-4-4-2 can disp12y the
entire message in one exposure; whereas, signs S-4-2-1 and S-4-2-2 require
tvrn exposures - ''SHARP TURN" follm.,ied by '1 NEXT RIDGE." Sign S-4-1-1 displays
the message in four one word exposures in sequence.
The four-word signs presented in Figure 1 and eight-word signs in
Figure 2 were designated by the sequential codes shown. An example will
be presented to illustrate the coding scheme:
S - Sequential display
4 - Words per message
S-4-2-2
2 - Words per sequential exposure
2 - Words per line 15
Example Message - 11 SHARP TURN NEXT RIDGE 11
SHARP TURN
S-4-2-2
~ SHARP TURN
S-4-1-1 S-4-2-1
SHARP TURN NEXT RIDGE
S-4-4-2
SHARP TURN
~ E
S-4-4-1
Code 2 divided by Code 3 gives the number of exposures required to present the entire message.
Figure 1. Four-word sequential signs displaying first exposure.
16
Example Message - "SHARP TURN NEXT RIDGE HHY-5 EAST SLOW DO~IN"
S-8-2-2
SHARP TURN! NEXT RIDGE HWY-5 EAST
1
SLOW om.JN i !
S-8-8-2
i
SHARP TURN ~ NEXT RIDGE
S-8-4-2
I
Figure 2. Eight-word sequential signs displaying first exposure
17
The eight-word sign display configurations, shown in Figure 2, operated
similarly to the four-word displays. The message visible at any particular
time of the eight-word signs S-8-2-2 and S-8-4-2 would be, in fact, identical
to the four-word signs S-4-2-2 and S-4-4-2, respectively.
Two run-on or moving message sign displays were also tested for both
four-word and eight-word messages. One run-on sign could display six char
acters, and the other ten. The smaller sign would be similar to sign S-4-1-1
in Figure and the larger sign similar to S-4-2-2. An example of a run-on
test code would be as follows:
R-8-2
R - Run-on
8 - Words per message
2 - Maximum word display
Experimental Development
The sequential sign displays presented in Figures l and 2 and the
run-on sign displays were simulated by motion picture photography using the
trailer-mounted matrix sign available during the laboratory studies. The
actual filming of the matrix sign was undertaken inside an aircraft ha~gar
that was converted into a photographic studio. This allowed control of the
surrounding interior light to a degree that the messa~es were in sharp con
trast to the darkened background, and thus clearly visible. A single line,
sequential display of a message was made by filming word groups of the message
statically displayed on the matrix sign. The word groups were selected to
provide the desired message sequence, and the film length of each group was
edited to provide the correct film frame count so that the message would be
18
shown for the correct time period. llo blank space was left between \'lord
groups. However, a blank space of 15 percent of the total message display
tise was used to define the end of a message for those studies having the
rn2~~2";e sequence initially starting in the middle of the message.
The development of the two- and four-line sequential displays required
a considerable amount of film editing and processing since the matrix sign
could display only one line at a time. A four-line sign display was pro
vided by adjusting the vertical angle of the camera four times so that the
message line being filmed would be developed in the proper position. Four
separate film strips, one for each line of the sign display, were edited
for correct time and then combined by photographic lab processing into one
film print. Two-line sign displays 1t1ere developed in a similar manner.
The filming of the run-on formats posed several different problems.
While the sequencing format could be filmed statically, and presentation
rates achieved by editing, the run-on formats had to be filmed in a dynanic
state directly from the matrix sign. When the sign was set at the higher
run-on rates, the punched tape reader exhibited a tendency to "skip."
Also, the messages v1ere of very poor visual quality due to a "flickeringll
effect caused by slow on-off ignition of the lamps. Neither characteristic
was evident at the slower run-on rate. In order to compensate for these
problems, the real-time, run-on speed was set at the slowest test rate
(l.O seconds/word). The camera speeds were then calculated and adjusted
to provide the other desired test rates. However, run-on word rates of
0.25 seconds/\·1ord still could not be provided with a high visual quality.
f..n r:1n-on formats were centered in the frame and filmed accordingly.
19
Four film strips were made. Two films presented messages beginning at
the first of a message and two began at the middle of the message, then
repeated the complete message. Tvm experimental texts, having randomly
selected messages, display type and display rates, were developed. These
same two texts were used for each message starting point. Each print con
tained four- and eight-word messages using both sequential and run-on displays.
Word messages and display orders were selected randomly from the given pool
of possible combinations while uniformly sampling all combinations. Two
samples for each sign and message display rate were shown, one in each cf
the two films for each message starting point. Within these constraints~
all experimental design factors were randomized or separated to reduce the
possibility of patterned or anticipatory response. The complete film scripts,
words, sign displays and display rates are presented in Appendix A.
Experimental Admi.nistration
The testing of subjects' ability to read and reproduce the displayed
messages, as simulated by observing the prepared motion picture films, was
conducted in the media-master laboratory previously described. The two
film strips having only messages beginning at the first of a message sequence
(films AF and BF in Appendix A) were viewed by 48 and 58 subjects, respectively.
The two film strips having messages beginning in the middle of a sequence
(AM and BM) were viewed by 70 and 50 subjects. No subject was used more than
once and never saw more than one film strip. Due to scheduling difficulties,
no more than five subjects ever viewed a film at one time. The subjects were
seated at tables before a large rear-projection photographic screen. The
subjects were supolied with answer sheets and were given taped audio
20
introductions concerning the study. They were shown slides of dynamic matrix
signs installed on a freeway as an introduction to the real-world situation.
I~structions were given along with a film example. The messages were then
sho',vn by means of a 16-mm projector with adequate time being a 11 owed between
mes:::ajes. The subjects were asked to completely re-construct each message in
writing. The instructions given and answer sheets used are included in
Appendix A.
Data Reduction
The critical. process in the reduction of this study data was to establish
a reasonable and consistent method of correct response determination. Exact
reconstruction of the message in the order presented was not required. A1so,
exact reproduction of route numerials in eight-word messages was not required.
The premise used in the consideration of a right or wrong response was whether
the message reported by the subject contained the major elements of the message
presented. The message could have been stated in a slightly different W3Y from
that presented, but it was scored correct if the meaning was clearly evident.
After some initial team evaluations, one researcher graded all responses. Upon
evaluating all subject data forms, average percentages of correct response ~·1ere
calculated for each presentation form, message length, display rate, and message
starting point.
Data Analysis
Analysis of the data provided considerable insight into human performance
with respect t0 matrix sign display characteristics. The data analysis revealed
subject limitatfons in reading and comprehension that may apply to all types of
changeable message signing systems. In general, word reading rates were similar
21
to those noted in the literature; however, subjects' ability to recall and
reproduce messages varied dramatically with message length.
The order of discussion of the results of the data analysis will consider
sequential signs first. Figure 1 depicted the four-word sequential sign
displays studied 111hereas Figure 2 presented the eight-word sign displays.
Following the discussion of results on sequential signs, run-on signing
display results will be considered. Performance associated with the remain
ing experimental design variables will be described for both sequential and
run-on signing displays.
Sequential Sign Results
The best performance for any sign display over a range of test conditions
was the two-line, two words per line sign shown in Figure 3. The sign was
displaying four-word messages beginning at the first of a message. At a word
display rate of 0.88 seconds per word, 96 percent of all subjects' written
responses as to the displayed messages were judged correct. The percent of
correct response was observed to drop as the display rate decreased~ but was
still measured at 79 percent correct response at a display rate of 0.25 seconds/
word. In cases such as this where the entire message is displayed at cne time,
the display rate is calculated from the time the full message is displayed.
The display rate would also correspond to the minimum possible subject reading
rate. That is, it would be possible for subjects to read faster than the
display rate, but not slower.
Display Rate - Except in a few instances, the capability of subjects to
correctly reproduce messages dropped as the display rate (seconds/word) de
creased, as illustrated in Figure 3. The few exceptions in this trend could
be attributed to some messages inadvertly being slightly more difficult to
22
w 100 CJ) z 0 a.. (/) LlJ a::
90
..... u w a:: 0:: 80 0 <..>
I-z w 70 (..) er: w a..
60 0.0
SHARP TURN NEXT RIDGE
S= Sequenced 4= Words per 4= Words per 2= Words per
02 0.4
DISPLAY RATE,
message exposure I ine
0.6 0.8
SECONDS/ WORD
Figure 3. Best four-word message sign display beginning at first of message.
23
1.0
read and recall than others, possible subject conditioning to certain message
presentation formats which either positively or negatively affected subsequent
test cases, deficiencies in the film prints, and possibly subject variability.
"r' ~ - t : ne e-;-tec s of display rate on performance will be further illustrated in the
following discussion of other results of the study.
0essage Length - The performance with the bJO-line, two words per line
sign presented in Figure 4 is typical of the performance results of all the
four-word messages tested as compared to eight-word ~essages. As· indicated
in Figure 4, the "best" average eight-word message response for all sign
displays, beginning at the first of the message, was only 40 percent at a
display rate of 0.88 seconds/word. At a display rate of 0.25 seconds/word,
no eight-word message display exceeded 10 percent. The studies show that
the eight-word messages having four "bits of information" could not be re-
produced by half of the test subjects. However, the eight-word messages all
included fictitious highway route numbers, which are more difficult to recall
than many common messages.
Number of Lines and ~ords per Line - Another issue studied was whether
it is better to arrange words horizontally in a display or to stack them
vertically. To test this issue, message length and number of sequences were
held constant. Only differences in words per line and number of lines were
of interest. The only direct comparison that could be made \vas between a one-
line, two words per line display (S-4-2-2) and a two-line, one word per line
display (S-4-2-1) each transmitting a four-word message in two sequential
exposures. The data were analyzed in terms of messages displayed at the first
and those starting at the middle.
24
w VJ z 0 a.. Cf)
w er:
t-(.)
w a::: 0::: 0 (.)
I-z w (.) a:: w a.
100
80
60
40
"0' "- .
SHARP TURN NEXT RIDGE
S-4-4-2
. SHARP TURN NEXT RIDGE +
II our-word S-4-4-2
HWY-5 EAST SLOW DOWN
S-8-4-2
II • 11 \ ""Eight-word
'--s-s-4-2
•-Best average response for al I signs
o~------_....--------"-----~------------------02 0.4 0.6 0.8 1.0
DISPLAY RATE, SECONDS/ WORD
Figure 4. Performance of same sign displaying four-word and eight-word messages from first of message.
25
The results of the two studies are shown in Figure 5. The results are
somewhat surprising. At display rates of 1.0 and 0.5 seconds per word, the
t\vo-1ine sign having one 1t10rd per line (S-4-2-1) performed better than the
one-line, two words per line sign (S-4-2-2). At 0.25 seconds/word, this
performance 1>ias true only for the message presented from the first.
A direct contrast 1>1as not made between a four-line, one-word per line
sign (S-4-4-1) and a one-line, four words per line sign since the latter sign
would have been difficult to simulate from the available matrix sign, and
probably is not a practical sign design to consider anyway. While the four
line, one word per line sign (S-4-4-1) performed well, it did not perform as
well in any test case as did the two-line, two words per line sign (S-4-4-2),
which was the best overall sign display. These results indicate that a square
message display provides an advantage over either a vertically or horizontally
elongated display configuration.
Number of Lines and Exposures - Subject performance to signs having one,
two and four lines were compared with contrasting sign designs having four,
two and one exposures. Four and eight-word messages beginning at the first
of a message were considered for three display rates. The results of this
comparison are presented in Table 4. The best sign display of the three for
display rates of 0.88 and 0.50 seconds/word was the t\-io-line by two-exposure
display for both four-word (S-4-2-1) and eight-word messages (S-8-4-2). At
the highest display rate, the two-line by two-exposure sign ranked second
when displaying four-word messages and third for eight-words.
A review of the results presented in Table 3 indicates that increasing
the number of lines of a display while correspondingly reducing the number
of sequential exnosures does not produce a consistent increase in performance.
26
w Cl) 2 0 a. (/) w a::
r-(.) LLl a: 0:: 0 u
r-z w u a:= w 0..
100
90
80
70
60
iii = S-4-2-1 SHARP TURN
A = S-4-2-2 I SHARP TURN I
-- Start at first --- Start at middle
JI
- ./6 /
/ /
/
A.- -
0 0.2 0.4 0.6 0.8 1.0
DISPLAY RATE, SECONDS/ WORD
Figure 5. Effectiveness of two-line, one word per line sign and one-line, two word per line sign.
27
Table 4. The effects of number of lines versus number of exposures over a range of display rates as measured by performance ranking.
Number Number Designation Performance Ranking of of of Display Rate (Sec./viord)
Lin es Exposures Sign 0.88 0.50 0.25
4 S-4-4-1 2 3 3
4 S-8-8-2 3 2 1
2 2 S-4-2-1
S-8-4-2
l
l
2
2 2
4
4
S-4-1-1
S-8-2-2
3
2
2
3
l* This sign has the highest average percent correct response for the three types of four-word signs listed in Table 3 for messages presented from first of message sequence.
l+ = This sign has the highest average percent correct response for the three types of eight-word signs listed in Table 3 for messages presented from first of message sequence.
28
3
2
For 1vord display rates of 0.88 and 0.50 seconds/word, the results indicate that
an average mix of two exposures and two lines to display a message is better
th~n extreme designs having either four lines and one exposure or four expo-
s~res and one line.
Startii\_9 Point - This particular factor was initially considered to eval
uate various sign display designs as to their relative performance when a
motorist (subject) might begin viewing a message sequence in the middle of a
message. The motorist was assumed to have to continue to read a subsequent
complete message sequence. For example, the S-8-2-2 sign (a one-line, b;o
~vords per line sign) would display the following eight-word message starting
at the middle as: 11 H~~Y-5 EAST", 11 SLOI~ DOWN 11, blank, "SHARP TURN", 11 NEXT
RIDGE'', "HWY-5 EAST", "SLOW DO\!JN".
A blank space was used to define the end of the message. This time was
set at fifteen percent of the total viewing time of the message, starting
at the middle point and measuring continuously until the message is completely
repeated. However, when this blank time exceeded 0.8 seconds, the blank time
was set to 0.8 seconds for reasons discussed below.
Two comments are offered concerning the duration of the blank space.
First, during some initial laboratory testing, subjects were too frequently
observed writing message responses to the simulated sign displays before they
had read the repeated message shown when the blank space on the film strip
exceeded aporoximately one second. One set of initial results confirmed
these observations. These problem cases were completely repeated using the
0.8 second blank time. Secondly, the blank time calculated from 15 percent
of the tota·1 viewing time, as experimentally shown, corresponds to 20 percent
of~ cycle time b~ing blank if a message was constantly being repeated or
cycled. 29
The evaluation of sign design performances included studies for both four
word and eight-word messages. Sign displays requiring one, two and four dis
crete sequential displays to transmit these message lengths were studied. It
was anticipated that sign displays requiring four exposures to display a given
message 1 ength would not perform as ~'lel 1 as a two- or a one-exposure display for
the same total viewing time. It was also expected that, for a given total view
ing time available to a subject (motorist), a drop in correct response to dis
played messages would occur where a subject began viewing a message in the middle
of a message sequence rather than at the beginning of the message. These expected
results were generally confirmed, with one noteworthy exception.
Figure 6 presents comparative results between messages displayed from
the first with those displayed from the middle for the same total viewing
time. Four-exposure and two-exposure sign displays are illustrated for
four-word and eight-word messages. For a given viewing time, viewing mes
sages from the middle was observed to reduce the performance with the signs
in six out of eight cases where comparisons are possible. No differences
occurred in the other two cases. For a given total display (vie\<1ing) time,
two-exposure signs out performed four-exposure signs in four out of five
cases. The results of Figure 6 are generally consistent with the results
expected.
The results from the one-exposure displays shown in Figure 7 were as
anticipated for the four-word message sign (S-4-4-2) but were surprisingly
different for the eight-word message sign (S-8-8-2) which displays the com
plete message at one time. For this latter sign, the complete message was
disnla.yed for approximately 28 percent of the total display time followed
by a 15 percent blank which is then followed by a repeat display of the
30
100
90
lLJ Cf) 80 z 0 Q_
(f) 70 w er:
-- START AT FIRST OF MESSAGE
--- START AT MIDDLE
-~ _ _. //-----.,. S-4-1-1 s -4-2-2
I- 60 II
FOUR-EXPOSURE II u 11
TWO -EXPOSURE II
w 0:: 50 0::: 0 u
40
z w 30 u er: w [)_ 20
S-8-2-2 S -P-4-2 ,_, I
I
I I
I
/
I
I 10 ;
I "w--- --- -ii ~
0 ~------...__..__....__.______..._ 0 2 4
SEC. 6 0 2
SEC.
TOTAL DISPLAY TIME
6 8
Figure 6. Effect of starting position and display time of message upon performance (two and four exposures).
31
w ({) z
100
90
80
0 70 0.. ({) w 0:: 60
.._ () 50 w er: 0::: 0 40 (.)
t- 30 z w (.)
0:: 20 w 0..
10
START AT FIRST OF MESSAGE
START AT MIDDLE
S-4-4-2
II ONE EXPOSURE II
S-8-8-2
__ -ii-------,. ·-
o __________________ _._ ________________ _ 0 4 8 12 16
TOTAL DISPLAY TIME - SECONDS
Figure 7. Effect of starting position and display time of message upon performance (one exposure).
32
complete message for 57 percent of the display time. For total display
times of 3.5 and 7.0 seconds, this on-off-on effect functioned about twice
as well as having the eight-word message displayed constantly for the same
display period. It was speculated that the flashing effect of the same mes
sage encouraged subjects to read the message twice while requiring perhaps
only half of the message to be remembered during each display. These results
further indicate that an eight-word message should be broken up by some means
or repeated so that motorists can effectively understand the intended message.
Run-On Sign Results
Run-on sign displays present a message as a train of words moving con
tinuously across a display area from right to left. Run-on sign displays
are also called moving message or continuous message displays. Two, cr.e-line
run-on display configurations v1ere tested in the laboratory studies. One
sign displayed six characters at one time and the other ten. It should be
noted that no word tested exceeded five characters for any run-on ~essage.
Display Rate - In all cases evaluated, the effectiveness of run-on dis-
plays, as meas~~2d by percent correct subject response to the messages pre-
sented, was reduced as the display rate in seconds per word was decreased.
These results for run-on displays shown in Figure 8 are similar to those
obtained for sequential sign displays. Most test res~1ts at the fastes~
display rate ( 0. 25 seconds/1·1ord) v1ere generally very poor. The messages
moved at such a high rate of speed across the display area that the words
were blurred even though they had been filmed from a slower, relatively
clear, moving message. It was not known whether better results could have
been obtained using an ideal matrix sign to display moving messages at 0.25
33
10()
w 80 U) z 0 CL U)
w 60 0:::
r-(.) LL.I 0::: 0::: 40 0 (.)
r-z w 20 (.)
0::: w CL
0 0
I SHARP I R-4-1
["SHARP TURN I R-4-2
R-4-2-.
R-4-1
t
R-8-2
R-8-1
0.2 0.4 0.6 0.8 1.0
DISPLAY RATE, SECONDS/ WORD
Figure 8. Effectiveness of six and ten character run-on displays starting at first of message.
34
seconds/word. Local TV experiences suggest that 0.42 seconds/word (12 char
acters/second) does not provide a generally acceptable reading rate, which
is supported by these study results.
>'.essa9e Len.9.0__ - fl.s Figure 8 shm·JS, subjects could read and reproduce
four-word messages much better than eight-word messages for a given display
rate. Four-word messages displayed from the first of the message at 0.88
seconds/word v-1ere reproduced with an accuracy exceeding 80 percent correct
response for both the one-word ( R-4-1) and two-word ( R-4-2) signs. Eight-
1tmrd messages could not be reproduced at an accuracy exceeding 16 percent.
Again. all eight-word messages required recall of route numbers which adjed
to their difficulty (See Appendix A).
\fords Per Line - The two signs tested \·1ere both one-line signs , :, .-'-,:,
i 11 ustrated in Figure 8. One sign could display six characters (about one
\vord) per 1 i ne, and the other sign could display ten characters (about tvo
words) per line. A slight advantage is possibly provided by the larger dis-
play (two-word) 1.;hen messages are vie1·Jed from the beginning. However, 1·1hen
four-word messages are viewed startinJ from the middle of the message t~e
larger display (R-4-2) has a clearer oerfornance advantage, as test results
presented in Figure 9 show.
R-4-2 sign for messages presented from both the first and middle of the
message. As might be expected, the message shown from the first resulted in
better subject performance. These results are s imi 1 ar to the four-sequence
and two-sequence sign results presented in Figure 6.
35
100_
~ 80 z 0 0.. (/) w 0:
lo w 0: 0: 0
60
(_) 40
1-z w (_) a:: w 20 0..
-- START AT FIRST OF MESSAGE --- START AT MIDDLE
R-4-2 R-4-2 _ ........ ---- R-4-1 -· -JIJ,,.---
1 -----I .--/ I
I I I I
j I
I I
I I •
I
R-8-2 __ -A A------- __ _. R-8-1 .....------
0--~--~--~--~-----------.._ ______ ~ 0 2 4 6 8
TOTAL DISPLAY TIME - SECONDS
Figure 9. Effectiveness of run-on signs for four and eight word messages starting in the middle.
36
Summary of Laboratory Results of Matrix Displays
A summary of all the signing displays tested in the laboratory from the
first of the message sequence are presented in Figure 10. It is clear that
rr:essage length is a very important factor in determining whether a message
is actually understood (could be reproduced). No display rate of eight-word
messages was tested for which the correct message response exceeded 40 percent.
Four-word messages, on the other hand, were read and reproduced by the
subjects to a high degree of accuracy at 0.88 secondsiword. Correct responses
from 90 to 95 percent were observed for the largest display rate tested (0.88
seconds/word). Over 85 percent of the subjects could read the more effective
sign displays at 0.5 seconds/word. At 0.25 seconds/word, over half of ~he
subjects could read and recall the four-'dOrd messages under laboratory con
diti ans.
One factor which might have p1ayed an important role in influencing
whether an eight-word message was read and subsequently reproduced by some
test subjects was whether they would initially try to read an eight-word
message in the first place, perhaps due to a personal awareness of their
own abilities a0d/ar existing motivation. It was observed during the studies
that some subjects appeared to be ove el~ed by the task of having to read
an eight-word message, knowing they would be asked to reproduce it. The un-
expected results of the dc::'.JL: :Jispl2yed, • ' I t ' '" e1gnt-worJ rnessaJes, presen~e~ 1n
Figure 7, show that a dramatic improverient in oerformance 1t1as obtained when
a long (eight-word) message was broken-up, either by being repeated or separ-
ated into sr.:aller ''chunks" (.§._) v1hich can then be read. Apparently, a four
word message (or sentence) is a message chunk size which motorists (subjects)
can read and understand efficiently. This premise is supported by the results
37
100
w 80 (/)
z 0 a.. VJ w a:: 60
t-(..)
w 0:::
40 er: 0 (..)
t-z w
20 (.) a:: w a..
0
S-4-4-2 S-4- 2-1 S-4-4-1 R-4-2 S-4-1-1 R-4-1
S-4-2-2
S-8-4-2
S-8-2-2
S-8-8-2 R-8-2 R-8-1
0.2 0.4 0.6 0.8 1.0
DISPLAY RATE, SECONDS/ WORD
Figure 10. Effectiveness of displays for four and eight word messages starting at first of message.
38
of the S-8-4-2 sign depicted in Figure 4. This sign separated eight-word mes-
sages into tvJO sequential di sp 1 ays of four-1110rds each in a two-line by tvJO
words per line format. Performance with this sign was the best of all eight-
word signs at 0.88 and 0.44 seconds/word, starting at the first of a message.
The second four words was a second sentence, the break serving as a period.
The laboratory study results suggest that the message 11 chunks 11 can be
more effectively read when the wording is displayed in a relatively compact
format both spatial"ly, in vertical and horizontal (as related to the nu111ber
of lines of display and words per line), and temporally (number of sequences
used to generate a chunk) but with each chunk having considerable separation
in time.
The most compact, four-word display is the two-line, two words per line
sign (S-4-4-2). As shown in Figure 3 (and 4), this sign design resulted in
the best performance for four-word messages (and also for eight-word ~essages
starting from the first). Perfomance results for th::i three sequential signs
presented in Table 4 (l line by 4 exposures, 2 lines by 2 exposures, and 4
lines by l exposure) further ind~cate that compact displays were generally
superior. In addition, signs having multiple exposures displaying message
chunks usually experienced a larger reduction in performance when subjects
began reading the sign 1 s message in the middle of the sequence, i.e., at
random.
The run-on message displays tested generally did not perform as Vieii
as did the sequential displays. However, the run-on displays did work
satisfactorily a~ t~e slower display word rate of 0.88 seconds/word (l.14
words/second). (H-h " v""en~1 se, the trends observed followed those of the
sequential signs.
39
FIELD EVALUATION OF MATRIX SIGNS
Introduction
Although the visual presentations shown in the laboratory studies were
motion picture replicas of an actual full scale matrix sign, it was not pos
sible to give the subjects the sensation of driving a motor vehicle. Exper
imentation in a "real world" environment such as on local major streets and
highways was deemed impractical for the procedures and experimental messages
to be used. However, the information to be gained from such experimentation
was felt to be quite valuable toward verifying the results of the laboratory
studies. Thus, a runway at the Texas A&M Research Annex was chosen as a
simulated highway for conducting "behind-the-wheel" studies using a new full
scale lamp matrix changeable message sign. The studies included both
"unloaded" experimentation in which the driver's movements, both laterally
and l ongitudi na lly, were virtually unrestricted and 11 1 oaded" experirnen~ati on
in \<Jhich the drivers \•1ere constrained later·ally by a very closely spaced
lane of cones.
Purpose
The p:·imary purpose of the field studies was to provide some verification
of the results of the laboratory studies. A secondary purpose of the field
studies was to obtain some indication of the legibi1ity distance character
istics of the lamp matrix sign. Another secondary purpose was to determine
whether longer exposure times to messages \•1ould improve driver performance.
The concept being explored was that when drivers are in congested traffic
(bumper-to-bumper) they have more time to read a sign and would thus be
exposed to the sign longer. The objectives of the field studies can be
40
s umma ri zed as fo 11 ows:
1 L. To determine whether subjects could perform as well (or better)
in an actual driving environment as in the lab.
2. To determine whether the rank order of the four-word message
formats would hold true for the field situation.
3. To determine whether the correct response rate for eight-word
messages would hold true under more realistic conditions.
4. To determine whether longer message exposure by use of a slower
word rate would improve correct response rates for both four
and eight-word messages.
5. To determine whether there was a substantial difference in subject
performance when the subjects were required to perform confined
lane tracking while reading the sign.
6. To make some estimation of the approximate legibility of the i8-
inch lamp matrix characters.
Scope
Although the previously stated objectives were broad based, the field
studies for the reduced scope was that the field studies were not intended
for development of new information, but rather for the verification and
refinement of the laboratory results, and for initial investigation of
hardware characteristics.
Matrix Sign
One of thr2e trailer-mounted lamp matrix signs purchased for use in this
project was used for the field studies (Figure 11). This computer driven
sicn contained two 7 x 64 arrays (lamp banks) of 33-watt bulbs generating
41
.=-
.. :..:- ~
Figure 11. New Trailer ~ounted Lamp Matrix Sign Used in Field Studies
42
alphanumeric characters 18 inches high. The lamp banks operated virtually
independent of one another and each could display messages from 0.25 to
511 seconds duration.
Exper::imental Design
As in the laboratory studies, the field evaluation study of matrix sign
displays could be considered a modified randomized factorial design. The
factors and the levels of treatment at which they were tested were:
Type - (1); sequential
Start - (1); beginning of message
Length - ( 2) ; four or eight words
Words - ( 2) ; one or two words per 1 ine
Lines - ( 2) ; one or two lines
Rate - (3); one-half, one, or two seconds per word
Loading - (2); unloaded, loaded
All studies were conducted once using a different message for each form2t
(sign design) and rate. Messages were randomly assigned to drivers to mini
mize the effect of any particular message on a sign test condition.
Four of the five, four-vwrd sign designs tested in the laboratory \-lere
tested in the field. All of the formats shown in Figure 1 were tested except
the S-4-4-1 format. This format was eliminated because it required the use
of a four-line sign and it did not appear to have any particular advantage
over the one and two-line designs tested in the laboratory.
Because cif the limited scope of the field studies, only one eight-word
sign design (S-8-4-2) was tested in the field. This format was by far the
bes:: eight-word format tested in the laboratory. This sign design was tested
43
under two configurations. The first was a sequential display similar to that
used in the laboratory of four words displayed for a given interval followed
im~ediately by the remaining four words in the message. The second configu
ration ·included a blank interval between the two four-word displays equal in
tir~;'~ to each of the message displays (i.e., two seconds on, two seconds blank,
two seconds on). A detailed description of the individual sign designs can be
found in the sect·ion on the laboratory studies. Two types of driver workload
studies were conducted - "unloaded" and 11 loaded 11• The 11 unloaded 11 studies
differed in concept from the laboratory studies only in that the ·subject i,.1as
behind the wheel instead of behind a laboratory desk. Thus, driving down a
150 foot wide runway with virtually no constraints to movement, the subjects
had virtually no workload other than reading the sign. In an effort to in
ject some additional workload in the 11 loaded 11 studies, subjects 1t1ere given
a lane of cones, Figure 12, which cleared the vehicle by about 6 inches on
both sides, in which to drive at the test speed of 45 miles per hour (35
miles per hour for eight-word displays). All subjects were tested for formats
and rates under both the "unloaded" and "loaded" conditions, with the "unloaded"
condition tested first.
The complete experimental design is presented in Appendix B.
Experimental Administration
As indicated previously, the field studies were conducted on a runway at
the Texas A&~ Research Annex. The subjects drove a 1976 Chevrolet sedan equip
ped with automatic transmission. As most of the subjects had driven the vehicle
during studies conducted as a part of another research project, little familiar
ization was required. They were accompanied in the vehicle by a test admini-
44
j;:L.,...· . . ...,_,,,.....
·,,:··"·
Figure 12. Vehicle Within Lane of Cones - "Loaded" Study
45
strator who, for purposes of consistency, was usually the same person who
administered the laboratory studies.
The administrator gave the subject the background and instructions for
the study and then asked the subject to make four preparatory runs. The
first three runs were made at approximately 20 miles-per-hour toward the
sign which displayed one of three legibility test words. These runs aided
in fami.liarizing the subject with the test environment and provided an assess
ment of legibility distance determined from distance ~arkers placed alongside
the runway. These legibility distances were used to ensure that test messages
v1ere not displayed before the subject v1as within his legibility dista.nce of
the sign. A final preparatory run was made at the test speed of 45 miles per
hour. A four-\vord message was displayed in a stationary S-4-4-2 format.
Subjects were asked to read the stationary message and recall it when they
had stopped past the sign. At that time they would repeat the message to
the administrator and return to the starti~g point to begin another run.
This procedure was followed throughout the studies. Instructions to the
subjects and answer sheets used are included in Appendix B.
Data Reduction
As in the laboratory studies, the purported concern in the reductio~ of
the field study data v;as not v1hether the subject cou1C: reproduce the r:;essage
exactly, but whether the message transmitted could be clearly interpreted
from whatever response was given by the subject. To maintain consistency
with the laboratory data reduction, the same individual \vas responsible for
determining the correctness of subject responses for the field studies. Upon
evaluating all subject data forms, percentages of correct response were
46
calculated for each sign design, message length, display rate, and loading
condition.
Reduction of legibility distance data consisted of determining the mean
cf t112 three legibility distances recorded for each subject. These means
v-1ere then transformed into measures of legibility in feet per inch of letter
height by dividing the legibility distances by 18 inches.
Data Analysis
There were four sign designs tested for the four-word messages, and one
sign design for the eight-word messages. Each design was tested at display
rates of 0.5, l.0, and 2.0 seconds per word. An additional design~ four-word
stationary in the S-4-4-2 format, was included for purposes of acclimating
the driver and for comparing stationary results with dynamic message test
data. The sign designs tested and the results of the studies are shown in
Table 5.
In an effort to inject an additional driver workload into J_: ~ ,
~ne s \..uay, a
narrow 1 ane of cones was formed for the 11 1 oaded" studies. Jl.lthough this ar-
rangement produced only a horizontal tracking vrnrkload, it was hoped that
this task would provide some simulation of freeway driving conditions. Cor
rect response rates for the four-word messages (Table 5) do not appear to
vary significantly bet\-Jeen the "unloaded" and 11 102ded!' conditions, indicating
that either the drivers were not loaded sufficiently by the tracking require-
ment, or that four-word messages can be read under virtually any loading
cor1 d·i ti on.
One of the primary functions of the field studies was to provide some
'Jeri fi cation of the results obtained in the laboratory s tu di es. For purposes
47
Table 5. Summary of Results of Field Studies
For Electronic Matrix Signs
--- Studii::ic: Field Field Laboratory Sign Display Unloaded Loaded Percent Display Rate Percent Percent correct Format sec/1;1ord correct correct
M
2.0 100 95 ---S-4-4-2 l. 0 100 l 00 96
0.5 90 100 93 I I
2.0 100 100 --- ! I
S-4-2-1 l. 0 95 95 95
0.5 80 90 I 86
2.0 100 95 ---
I S-4-1-1 l. 0 100 100 88
0.5 75 85 85
2.0 100 100 --- I S-4-2-2 l. 0 100 100 78 I
0.5 80 80 74
S-8-4-2 2.0 70 55 ---w/o Blank 1. 0 55 40 40
S-8-4-2 I I 1,1/Bl ank 0.5 40 10 22 i
I Stationary 100 95 i --- --- I
48
of comparison, the related laboratory results are shown in the last column
of Table 5. The results of the field studies in general confirmed many of
the findings of the laboratory studies, although there was a marked improve-
1~ent in some areas. A fairly representative assessment of the relationship
bet1'i::en the laboratory and field results may be obtained from the composite
comparison of the S-4-4-2 format shown in Figure 13. Regardless of loading,
the subjects performed well on the sign reading at the slower display rates
with some degraded performance at the faster rate (0.5 seconds/word). The
field results for this sign design were fairly typical for all of the designs
in that performance was some~<1hat higher than the laboratory results at the
slower display rates (l.O seconds/word) and equal to or bounding the labora
tory results at the faster display rate (0.5 secondS/\·1ord).
None of the field study results were different by more than one correct
response (five percent) for the display rates of 1.0 and 2.0 seconds/word.
The fact that at these display rates, the correct response rate w2s 95 per
cent or better indicates that the reading/recalling task was not very dif
ficult. However, at 0.5 seconds/word, the percentages of correct response
vary considerably. It is possible that s~1bject response to a sign display
at this faster rate may be the best measure of the display 1 s capability.
Thus, correct response at 0.5 seconds per word was used to rank order the
four-word formats for each sign display.
Probably one of the more significant results of the field studies was
the confirmation of the rank order of the four-word sign designs obtained in
the ·1 abora to l'Y s tu di es. The four-~<1ord fornat performance scores shown in
Table 4 were presented in the rank order according to the laboratory results
for the 0.5 s<.::condsfi,-mrd display rate under the 11 loaded" condition. The
49
w (f) z
100
~90 (f) w CL
I- 80 u w 0::: 0:::
8 70
1-z w ~ 60 w CL
I
j
I
I S-4-4-2
SHARP TURN NEXT RIDGE
LEGEND
m- - - LABORATORY
® FIELD UNLOADED
A--- FIELD LOADED
50.__~__,_~~--'-~~-'-~~~ 0.0 0.5 1.0 1.5 2.0
DISPLAY RATE, SECONDS/WORD
Figure 13. Comparison of Field and Laboratory Results for a Selected Sign Display.
50
percentage of correct response for signs S-4-4-2, S-4-2-1, S-4-1-1, and
S-4-2-2 was 100, 90, 85, and 80 respectively. Table 6 shows a simple
rank ordering from 1 to 4 for the formats tested in the laboratory, field
u~l:2Jed, and field loaded studies. Although there is some variation in the
r:rn;:ing of the third and fourth best designs, the ranking of the first and
second best sign displays is consistent within the three studies.
Although a practical maximum fast display rate of 0.5 seconds per word was
fairly well established in the laboratory, there was little indication of \I/hat
a minimum slow display rate should be. One boundary for minimum word display
rate is of necessity the rate at which an entire message can be presented at
freeway speed within the average legibility distance of drivers. That rate
was found to be about 2.0 seconds/word. This rate was included in the field
studies to determine \'Jhether its use \vould improve the correct response rate.
The data from the four-word messages do not indicate that the 2.0 seconds/word
disolay rate is better than the 1.0 seconds/word rate.
Responses to the eight-word sign design suggested that this farsat je
discussed separately. There was some concern as to whether the laboratory
results for Lhe eight-word messages were representative of actual conditions.
The field study results indicate that, whether or not subject motivation in
the laboratory v;as actually a factor, response to the eight-word messages was
considerably better than in the laboratory. \·lith regard to the disp1c.y rates
used, the ei ~~ht-word messages showed cons i derab i e improvement in correct
response rate for the 1.0 seconds/word display rate (Table 5). Questionable
imprnvement in correct response occurred at 0.5 seconds/word as the loaded
shdy performance was poorer than the corresponding 1 aboratory performance.
Res ~o'1ses to the eight-word messages shm·1ed a substantial difference between
51
Format ... S-4-4-2
S-4-2-1
S-4-1-1
S-4-2-2
Tab 1 e 6 : i
Rank Order of Four-Word Sign Designs at 0.5 Seconds per Word Rate
Rankina Field Field
Unloaded Loaded
l 1
2 2
4 3
3 4
Laboratory
1
2
3 i
4 I
''unloaded" and "loaded" conditions at all three diso1av rates. These de.ta ' _,
indicate that correct response rates for eight-word messages are very
sensitive to the level of driver workload encountered.
Lamp Matrix Sign Legibility
The final objective of the field studies was to estimate the legibility
of messages generated by the 18-inch character lamp matrix sign. The
IES Lighting Handbook (§_) recommends that letter height be computed as:
D Hr = 500 '
where: Hr= minimum letter height (feet)
D = maximum distance at which letter is legible to a majority of people" (feet).
This equation suggests an assumed legibility distance of 41. 7 ft/inch of
letter height. Bogdanoff and Thompson (Z) concluded from an undocumented
field study that an 18-inch character lamp matrix sign was readable at 800 ft
to the "average motorist." This approximate distance translates to about
44 ft/inch of letter height.
..
While these sources did not provide definitive supporting data, they
did serve as a basis for comparison. Further, it was felt that not only
shou1d the "average" or "majority" legibility distances be investigated,
but the 85th percentile as we,11, as it v10ul d probably more nearly represent
a 11 c!2s·ign'' value.
The field study consisted of a determination of the maximum distance
at which each of 20 subjects could read a test word. Each subject read
three test words, randomized in order between subjects. The test words used
were "BOAT, 11 "BOOK," and 11 ROCK. 11 These words were chosen because they had
been shown to be of very nearly equal legibility in a previous TTI study (.?_).
The mean legibility distance of the three trials was used as the subject's
legibility distance in further computations.
Each of the twenty subjects used \I/as a licensed driver with a known
corrected static visual acuity. Subjects were chosen from a subject pool
to as replicate, as closely as practical, a national cross-section of drivers.
As these subjects were used in several other field experiments during the
testing period, corrected static visual acuity was only one of several
selection criteria. HovJever, as sho•:m in Figure 14, their measured visual
acuities were fairly close to that of the national driving population (1).
The results of the study indicated that the previous estimates of
legibility of lamp matrix signs were fairly accurate with respect to the
"average" driver. Hov1ever, it apoears that a IT'ore conservative estimate of
legibility distance may be in order for design purposes .
As each of the test words had been shown to be fairly equal in legibi-
1 ity, the mean of the three distances was computed for each subject. These
mean::; •,.1ere t·1e-r plotted on a cumulati··e distribution (Figure 15).
53
100
90
S-
~ 80 +-' (!)
ro So s:: 70 3 0
...s::: If)
~ .,.... 60 :::i u < ,..... rd :::i VJ 50 .,....
>
.,.... 3
+-' s:: <lJ u S(!)
0...
ClJ >
·.-
::I
40
30
E 20 ::I u
10
----'3 Subject Populaticn
m:111!11eaaar11Q National Population
20/10 20/20 20/30 20/40
Corrected Static Visual Acuity
Figure 14 Comparison of Visual Acuities of Subjects With National Population
54
70 60 50 40 I .(ft./"! r1. )
,./\ -;.._ 900 800 700 fiQO ' I I v 1200 1100 l 000 I I I I I I l ft. J loo A,. I I I I
90 I - - - - - - - - - - --
80 1
w u I c 70 Ill ..._, Vl ....
Cl
>, 60 ..._, -~
~ .... ..a I- 50 .... w O'> ..._,
OJ "' ....J OJ L..
..C:. l.!l 40 U1 ..._, U1 .... I-
:I 0 ..._, c c ): 30 'OJ 0 u ..c: I- VI Q.!
Q.
Q.! 20 > .,..
...... II) r-
10 :::i s u
Q...L../ (m) 3so 300 I 250 200
~ {m/cm) 8 7 6 5 4
Legibility Distance
Figure l'.:i Legibility Distance Observat'ions
The mean legibility distance for all subjects was about 840 ft. This
distance translates to 46.7 ft/inch of letter height. From Figure 15 the
median legibility distance was 860 ft, and the 85th percentile, 637 ft.
These distances translate to 47.8 ft/inch and 35.4 ft/inch, respectively.
The ana'lysis showed the 85th percentile lebibility of distance of a
18-inch matrix sign to be about 35 ft/inch of letter height. The closeness
of the corrected static visual acuities of the subject population and the
national population and the general agreement between the study mean and
previously reported averages further substantiate the results. As the
study considered only one size of lamp matrix display, it is not possible
to generalize the reported legibility distances to other sizes or types of
matrix displays.
56
LABORATORY EVALUATION OF LAMP MATRIX BULB LOSS
Introduction
The objective of motorist information systems, whether audio, visual,
static, or dynamic, is to transfer meaningful messages to the motoring public.
These messages may pertain to various tasks associated with vehicular maneuvers
such as route guidance, traffic conditions, or hazard warning. In displaying
information by variable matrix electronic signs, the legibility of the \'lords
displayed is the critical first step in message transfer. A designated portion
of motorists must be able to effectively read the words shown. If the dynamic
display fails in this capacity, then the display is useless and message trans-
fer cannot be achieved.
In the operational setting of an electronic display, it is reasonable to
expect that one or more matrix sign bulbs may be 1ost and drivers are required
to read the sign before the bulbs can be replaced. The experimental question
was: "How much bulb loss is tolerable before a message is misunderstood or
misinterpreted?" The emphasis in this study was to measure human comprehension
of traffic-condition words or route numerals of various lengths as displayed on
a variable matrix sign under various conditions of bu1b loss. The percent of
bulb loss at which alphanumeric messages become illegible was determined.
Specifications for bulb replacement ~ere recommended based on these established
bulb loss percentages.
Experimental Deyelopment
The sing.le 1ine lamp matrix sign, used in all laboratory studies, was
described earlier. The physical dimensions of this equipment imposed an upper
l i::·ii t on \vOrd 12ngth to ten characters. Four character words were chosen as
57
the lower limit. Words of fewer characters are generally prepositions, con-
junctions, and adjectives and were not considered, as the interest within this
study was primarily with one and two word combinations. Five different sets
of "Highway Condi ti on 11 v10rds \vere chosen for each word 1 ength. The independent
vari ab 1 e was length of words varying from 4 to 10 characters. The word list is
shown in Table 7. Five different route numerals were also chosen, for a total
of 40 words and numerals (5 X 7 + 5 = 40). These were subsequently divided I
into two groups or sets of 20 each (Tab 1 e 7) .;
The individual word or numeral was to be presented statically on the
electronic matrix sign with various degrees of bulb failure to be simulated.
Initial observations indicated that virtually no unfamiliar words were
legible beyond a 50 percent bulb loss. An unfamiliar word is defined, within
this study, as a word which has not been recognized and read at a 1esser degree
bulb loss; a familiar word exhibits the opposite characteristic. Therefore,
five equal increments of 10 percent bulb loss were established ranging from
10 percent to 50 percent inclusive.
As there is no real-world pattern to bulb loss, it was desired to simulate
random bulb failure. A chart was plotted duplicating the actual 7 X 60 matrix
exhibited by the sign. By using column and row assignment within the ~atrix,
random bulb "failures" were generated from random number tables. This process
was continued until 42 positions, or 10 percent of 420 bulbs (7 X 60), were
selected to simulate bulb failure. The corresponding bulbs were turned off by
simply unscrewing the selected bulbs. Each word or numeral was then displayed
on the sign and 16-rrrn slides photographed. The same procedures were repeated
for all percentages of bulb loss (10, 20, 30, 40 and 50 percent).
58
TABLE 7 Word Lists Used in Bulb Loss Study
CH,~RiKTER LENGTH GROLJD 1'1 GROUP B ,..,
4 Sl O'.'i Lane Toll Exit Road
5 Truck Alert Route ~freck
Merge
6 Bypass Bridge Access Median Reduce
7 Blocked Freeway Traffic Sta 11 ed
Vehicle
8 Accident Downtown Entrance Junction Pavement
9 Condition Diversion Alternate Hazardous
Collis-iJn
10 Congestion Restricted Expressway Prohibited Visibility
Route Numerals I-il - HvlY-6 -,
~
US- 3 I-270 US-39
59
The slides of the 11 Highway Descriptor" words and route numerals at the
designated degrees of bulb failure were arranged randomly within each of the
tv10 groups, "A" and "B". Each group 1vas then arranged further into t1t10 series
of presentations; an increase from 10 percent to 50 percent bulb loss and a
decrease from 50 percent to 10 percent bulb loss. Each complete order and
series are 9iven in Appendix C.
Increasing loss (10 to 50 percent) was assumed to represent a condition
experienced by a familiar driver when the word or numeral is seen clearly
legible and then gradually degraded over time until recognition is not possible.
Decreasing bulb loss (50 to 10 percent) was designed to test unfamiliar drivers
viewing the sign for the first time. Not having seen the word before, the sub
ject must gradually perceive the word by piecing together the elements. Thus,
the familiar series involved recognition only of previously intact ~ords while
the unfamiliar series involved a gradual process of grasping the meaning.
Each series was measured separately to obtain the performance of the
familiar and unfamiliar drivers. Averaging the ascending and descendi~a series
according to the psychophysical method of limits, it was also possible to off ..
set errors of anticipation with errors of perseveration and obtain an average
value which was most representative of driver recognition.
In summary, the independent variables investigated \vere as follO'-'IS:
8 Character lengths
5 Different words or numbers per character length
5 Increments of bulb failure (Random)
40 ~iords/:L.mera ls per study order
2 Series of presentation per order
(!) 50 percent to 10 percent bulb loss (unfamiliar)
60
(2) 10 percent to 50 percent bulb loss (familiar)
Further details concerning the experimental design of this study are given
"in Appendix C.
fls in thE~ first laboratory study, subjects \'Jere tested in the media-
master laboratory. All subjects selected for this study were Bryan/
College Station residents and drawn from the population pool of Table l.
The tot a 1 number of subjects tested was 93, divided as follows:
GROUP PRESENTATION SERIES NUMBER OF SUBJECTS
A 50% - l 0% 26
B 50% - 10% 25
c l 0% - 50% 25
D 10% - 50% 17
93
Each group of subjects was administered 100 words - 20 for each level
of bulb loss. The words were given in a different random order at each bulb
loss level. From one to five subjects were tested at any given time. Taped
voice instructions were played to the subjects and an example slide was dis-
played onto the opaque wall screen using the rear-projection method. Each
word or numeral, with a given percentage of bulb loss, was then projected
upon the screen for a 3-second period of time. The time of slide display was
chosen to be J reasonable value of which a driver would have visual exposure
as approaching a sign of standard legibility design at a normal operating
speed. The sl~de was then removed from the screen, and the subjects were
required to completely and legibly write the word or numeral if such was
61
discernib1e. Ten seconds were given for response before the next slide was
presented. This was found to be ample time for a written response. The
study script and response forms are shown in Appendix C.
Data qeduction and Analvsis -----·---------- ""
The criteria chosen for correct response to bulb loss was that the sub-
jects must complete and exactly reproduce the word or numeral displayed. In
other words, an incorrect response, or error, was recorded if the subject
either omitted the word (numeral) completely or reproduction was incorrect.
The following formula was used to calculate percentage correct response for
a given word length and percentage bulb failure:
[l - E/NJ • 100%
where
E = Total of errors - either by omission or incorrect reproduction
N = Number of words presented at a designated bulb loss and of a
designated word length.
As discussed previously, each group of words and numerals was analyzed
for the ascending series from 10 percent to 50 percent, and a descending
series from 50 percent to 10 percent bulb loss. During the data reduction
process, the two series were not only analyzed separately, representing the
familiar and unfamiliar motorist condition, but were also evaluated in total.
Summaries of percentage bulb loss versus percentage correct response versus
word length are given in Table 8.
Figures 16 and 17,depict the curve plots for the familiar motorist
bulb failure ascending from 10 - 50 percent; and the unfamiliar motorist
bulb failure descending from 50 - 10 percent. The unfamiliar driver series
represents the "worst con di ti on. 11 The average, total of both series, is
62
Table 8 Summary of Percentage Bulb Loss Versus Percentage Correct Response Versus Word Length
CHARACTERS % BULB MOTORIST CONDITION
PER \·JORD LOSS FAM IL I.l\R UNFAMILIAR AVERAGE
4 C/W 10 99 96 96 20 98 91 94 30 97 79 88 40 91 51 69 50 80 29 54
5 C/W 10 l 00 97 99 20 99 92 96 30 98 77 87 40 92 37 61 50 79 10 42
6 C/W 10 98 94 96 20 98 78 88 30 96 57 76 40 91 29 59 50 79 13 46
7 C/W 10 100 94 97 20 99 89 94 30 98 82 90 40 93 63 76 50 83 . 28 47
8 C/\-1 10 98 93 95 20 96 80 84 30 93 65 78 40 86 38 62 50 78 9 44
9 C/W 10 97 88 91 20 96 70 84 30 95 50 74 40 88 26 54 50 82 7 40
10 C/W 10 99 92 95 20 98 78 88 30 95 59 77 40 85 34 55 50 78 12 49
Numerals 10 100 94 97 20 97 90 93 30 96 70 82 40 84 40 60 50 59 8 35
63
w ({)
z 0 CL ({) w 0::
t-o w n:::
100
90
5/c, 7 /c1
4/c 1 10/c
6/c, 8/c
9/c
-~
~5th Percentile_
7/c
gs 80 0
8/c, 10/c
70
10 20 30 40 50 0/o BULB LOSS
Figure 16 5J1b Loss Versus Percent Correct Response As A Function of ~lord Length in a "Familiar" Motorist Condition
64
40
10 20 30 40 50
0/o BULB LOSS
Figure 17. Bulb Loss Versus Percent Correct Response As A Function of Word Length h an "Unfamiliar" Motorist Condition.
65
shown in Figure 18. Lines were drawn on all_ graphs at the 85th and 95th
percent levels of correct response. These references are used commonly in
traffic engineering practice as the basis for design recommendations.
The data indicates the following:
(1) At all levels of bulb loss, the unfamiliar series resulted in
poorer recognition than the familiar series. This was expected
since the subject had previously seen the words under low 1oss
levels and hence, needed fewer parts of the words in order to
recognize them under higher levels of bulb loss.
(2) The length of the word bore no systematic relationship to
percentage correct response for the familiar series (Figure 16),
but for the unfamiliar series the longer words were somewhat
more di ffi cult to recognize than the shorter 1;wrds (Figure 17).
As can be readily seen from the graphs in Figures 16, 17, and 18, the
curves do not indicate a relationship between word length and performances.
At first the finding seemed to be inconsistent with other studies sf word
recognition which indicate that 'tJOrds of a greater number of characters are
read at a higher percentage of degradation than words of a lesser number of
characters. For example, 11 Affirmative" is more easily recognized than "Yes".
It was suggested that the better recognition of shorter words may be due to
a higher frequency of occurrence of shorter words in the "Traffic Voca.bul ary"
than that of the longer words used. In an attempt to test this suggestion,
the frequency of occurrence of traffic words of varying lengths was deterr;iined.
These ~vords v;ere se 1 ected as being typi ca 1 of those di sp 1 ayed on dynamic motor
ist information signs through the U.S.(_!).
66
w (()
z 0 a_ CJ) w 0:::
I-u w 0::: 0::: 0 0
~ 0
100 5/c 7/c
Numerals, 4/c _95th Percentile 6/c
I c 8/c
90
a.5!._h P~entile_
80
70
60
50
40
10 20 30 40 50 0io BULB LOSS
Figure 18. Bulb Loss Versus Percent Correct Response as a Function of \,ford Length in an "Average" Motorist Condition.
67
Results
From the preceding figures, bulb loss percentages as associated with the
criterion perfor-mance 1 evel s of 85 and 95 percent were graphically determined.
These percentages, for 4 to 10 character words and route numerals, were
surcnarized for all conditions (Fariiliar, Unfamiliar) and presented in Table 9.
Many different viewpoints can be taken in arriving at conclusions in
this study. For the freeway corrmuter or fami1iar driver, approximately 45
percent and 30 percent bulb loss, corresponding to the 85th and 95th
percentile design levels, is tolerable before deterioration reaches a point
where legibility is a problem. Of course, poor appearance, and possible 1oss
of credibility may merit bulb replacement before this level of loss develcps.
The condition involving the unfamiliar motorist will only allow taler-
able bulb failure percentages of approximately 20 and 10 percent for .,_. :..ne
corresponding design levels of 85 and 95 percent. These loss percentages
are more consistent with appearance criteria and suggested manufacturer 1 s
bulb replacement specifications of 10 percent loss (10). Also, the dependence
of the unfamiliar driver on dynamic signing information is an argument in
favor of the 10 percent criteria. The tolerable bulb loss percentages of
approximately 30 percent (85th percentile) and 15 percent (95th percentile),
as shown for the average condition, are possibly more representative of a
normal stratification of familiar and unfamiliar motorists within the driving
public. However, at the 85th percentile design levels, appearance would be
questionable; thus, a 10 percent bulb loss is recommended. Further, it
should be emoho.si zed that these results :·1ere based on laboratory studies
in which the subject could devote his comolete attention to the displays.
In an actual driving situation the driver could not be expected to devote
68
---
Table 9. Bulb Loss Percentage Associated With Criterion Performances of 85% and 95% as a Function of Word Length
NUMBER PERCENT BULB LOSS OF
CHARACTERS FAMILIAR UNFAMILIAR. AVERAGE
85% 95% B'""C/ ::> ;o 95% s5;£ 95%
4 45 28 23 11 31 17
5 44 31 21 12 32 21
6 43 25 20 8 31 17
7 49 31 21 10 30 20
8 41 25 20 6* 25 6*
9 46 23* 10* 8 16* 7
10 42 28 15 7 28 10
ROUTE ilUMERALS 36* 31 F 8 26 17
AVERAGE 44;~ ?"'' ~0.'0 ~-l " 8'; 2 s;~ 14 ';
* - Minimum Bulb Loss Tolerable for Criterion Performance regardless of Length
69
I I I I I I
I l
I
as much attention to the displays. Thus, it is probable that he could not
tolerate as much bulb loss as indicated by the laboratory studies.
As most dynamic motorist information systems display messages involving
twJ word combinations on one line, this fact must al~o be taken into
consideration. A message may consist of two words of different lengths with
different percentages of bulb loss tolerable for legibility. The poorest
performance measured by driver recognition of a word of a specified character
length becomes the critical factor in message transfer. In Table 9, ,the
bulb loss percentages shown by asterisks are the smallest losses allowing
criterion performance regardless of \'/Ord length for familiar, unfamii1ar,
and average drivers. These percentages, rather than average values, set
standards when messages have mixed or unknown lengths.
Route numeral performance, from a bulb failure versus legibility stand
point, was 36 percent at the 85th percentile design level, while nine char
acter words performed at 23 percent for the 95th percentile design level under
the familiar condition. Corresponding lowest bulb loss percentages were 10
percent and 6 percent by nine and eight character words for the unfamiliar
condition and 16 percent and 6 percent also by nine and ei~ht character words
under average conditions. These performances should be considered in bulb
replacement for multi-word messages.
As indicated by the data, route numerals pose special problems of concern
1-1i th degrade.ti on and l egi bi 1 ity. Unsatisfactory performance, under average
conditions, is exhibited for the 85th percentile correct response level beyond
an approximate 20 percent bulb failure and for the 95th percentile level beyond
an approximate 10 percent bulb loss. This indicates that tolerable bulb loss
criteria for both legibility and appearance of route numerals are closely
70
I I
related. Special bulb specifications should possibly be considered when
using messages with route numerals. The literature indicates that numbers
are harder to recognize than words because there is no "sequential redundancy, 11
i.e., knowing one number one cannot anticipate the next; while the language
of words does permit filling in missing or distorted letters.
Conclusions and Recorrnnendations
Severa 1 con cl us ions and recorrnnendations concerning the effects of
bulb loss on the legibility of words, route numerals, and messages displayed
on variable matrix electronic signs are suggested by the results of this
study. Some are as follows:
(1) For 85% or 95% of traffic-related words to be correctly read,
the percentage of bulb failures must not be greater than indicated below:
MOTORIST CONDITION PERFORMANCE CRITERIA (:i BULB LOSS)
85th 95th
FAM I LIAR 44 28
UNFAMILIAR 18 8
.CW ERA GE 28 14
(2) Bulb replacement criteria for a specified level of legibili~y per-
formance vary with the motorist condition as indicated above.
( ') \ Jj Legibility perfon~ance under degradation due tc bulb loss is
dependent upon 1,;1ord length. 1-Jith the unfamiliar motorist, shorter words
\'/ere easier to recognize than longer \'lords at every level of bulb loss. It
is suggested that this may have been due to the shorter words being more
common in the language.
71
(4) At the 85th percentile performance criteria, for both familiar and
unfamiliar motorists, bulb replacement may be controlled by appearance rather
than legibility. The matrix sign may be legible at a level of bulb loss at
which the overall appearance might be unacceptable.
(5) Only in the unfamiliar case and at the 95th percentile does the
replacement bulb loss percentage obtained in the laboratory studies approach
that designated by sign manufacturers (Approximately 10%).
(6) In multi-word messages, the word length exhibiting the worst per
formance controls bulb replacement specifications.
(7) Messages with route numbers are read with difficulty at bulb failures
beyond approximately 15 percent. Special considerations are advised for route
numeral bulb replacement specifications.
(8) There is need for further study to evaluate more completeiy
legibility performance with route numerals and three character words under
bulb loss degradation.
72
LABORATORY EVALUATION OF SYMBOLIC MATRIX SUBSTITUTION
Introduction
The diversion of drivers from a primary route to and along an alternate
ro~~ poses some unique problems with respect to lamp matrix changeable mes-
sage signs. As it is likely that some of the diverted drivers are unfamiliar
with the alternate route, it may be necessary to provide guidance along the
route. This guidance would likely take the form of trailblazers. To reduce
the confusion associated with trailblazers for named or numbered routes, the
feasibility of using symbolic trailblazers are being explored in another
phase of this research project. Using this technique, the lamp matrix change
able message sign could display 11 ••• FOLLO~J (symbol). 11 Symbolic trailblazers
could then be placed along the alternate route to guide the diverted motorist
to his destination or back to his primary route.
However, symbols formed on a matrix sign are not necessarily exact r2p-
licas of the same symbol formed on a painted sign. Thus, the objective of
this study was to "evaluate the ability of subjects to associate correctly
symbols formed on a matrix sign with corresponding symbols formed on a
painted static sign (trailblazer)."
To accomplish this objective, several of the most readily recognized sy~-
bols were assimilated from available human factors research (11). These symbols,
shown in Figure 19, were reported to be highly discriminable from each other.
Each of the available symbols was systematically evaluated so that none
violated any of three criteria:
1. The sy~bol should be easily constructed on a matrix sign.
2. The sy~bol should not be confused easily with symbols used for other
high~vay signs.
73
FIGURE 19. HIGHLY RECOGNIZABLE SYMBOLS (11._)
74
3. The symbol should be readily acceptable to the general public.
These evaluations resulted in the elimination of several of the 15
symbols shown. The airplane, the crescent moon, the six-point star, the
heart, and the bullet were eliminated because they could not be constructed
easily on the matrix sign. The cross was eliminated because of its simi~
larity to a standard highway cross road sign. The swastika was eliminated
because of its obvious unacceptability to the American public. The remain
ing eight symbols and variations thereof were then programmed into the exper
imental design.
Experimental Design
The experimental design of this laboratory study was based on a frequency
count of multiple choice responses to various symbols. Each subject was briefly
shown a matrix symbol and then four painted symbols, of which he was to select
the one that most resembled the matrix symbol. Twelve matrix symbo1s were
shown to each subject. To sinulate the brief exposure to the symbols that
the subject would most likely encounter in the driving environment, the expo
sure of the matrix symbol was limited to three seconds, and to the painted
symbols, five seconds. Thus +:ne subject did not have time to "study" the
painted symbols.
Experimental Develo~nent
The matrix sign symbols shown in Figures 20-24 were simulated by still
slide photography of the one-line matrix sign used in the laboratory studies.
The painted symbols adjacent to the matrix symbols were prepared by the
staff artist according to dimensional specifications dictated by the 7 X 7
matrix lamp configuration.
75
A B c D
B c D
Symbol I-3
CJ . -
A B D
Symbol I-4
D . -
A B c D
FIGURE 20. GROUP I SYMBOLS - SMOOTHED CURVED LINE FORMS
76
S-v1nbol II -1 ___ _._
A B c D
Symbol II-2
A B c D
FIGURE 21
GROUP I I SYMBOLS - OUTLINE FORMS
77
Symbol II I -1
# '
. : ,
A B c D
Symbol I II - 2
* A 8 c D
FIGURE 22 GROUP III SYMBOLS - LINE OR SNOWFLAKE FORMS
78
Symbol IV-1
DD A B c D
A B c D
Symbol IV-3
L A 8 c D
FIGURE 211 GROUP IV SYMBOLS - OTHER FORMS
79
1J
Symbo 1 IV - 4
~~
GOW == Green on vlJhi te
BOW== Black on White
GOW y BOW
FIGURE 24
BOW
GROUP IV SUMBOLS (CONT.) - COLOR SUBSTITUTION
80
GOW
Development of the symbols used began with the eight readily recognizable
sy~bols mentioned previously. As it was not possible to replicate exactly
many of the symbols on the 7 X 7 matrix, a nearest approximation was formed
on t~e matrix sign. The painted symbols were then designed to include the
ori,;Jinal symbol, a literal interpretation of the matrix symbol (if it differed
from the original symbol), and other similar symbols to provide a choice of
four.
Although the symbols were presented random1y i:1 the 1aboratory studies,
they are purposely arranged into four groups in this report for discussion.
Group I (Figure 20) consists mainly of symbols whose original form contained
smooth curved 1ines. However, as indicated previously, they had to be slight1y
modified for display on the matrix sign. Thus, the literal formation of the
matrix symbol was somewhat different than the original .form. For all of
symbols, the original symbo1 is outlined with a solid box, and the li~era1
formation of the matrix symbol is outlined with a dashed box where applicable.
Group II symbols were those that consisted of outline forms (Figure 21).
In both cases, there were two painted symbols that correspond to the general
shape of the matrix symbol, one of 'dhich 1·;as out of prooortion. G,...c;ups III
(Figure 22) contains line or snowflake symbols that were generated because of
their ease of construction on a matrix sign and their lack of similarity to
other highway symbols. The symbols found in Group IV (Figures 23 and 24) ~ere
so grouped because they included such factors as line weight, negative repro-
duction of the matrix symbol, and color substitution. The color substitution
studies were ·included to detennine v1hether color had any effect on subject
selection of painted symbols.
81
The 11 pairs of black and white symbols slides were arranged in random
order behind a pair of example slides (Figure 25). The color substitution
symbols were developed and tested subsequent to the original study, and thus
were administered with only the example slides.
Exoerimenta~ Administration -~------
The testing of subjects' ability to associate correctly symbols formed
on a matrix sign with corresponding symbols formed on a painted static sign
was conducted in the media-master laboratory previously described. The 11
original symbols were viewed by 50 subjects, all seeing the symbols in the
same random order (See Appendix D). No more than five subjects viewe~ .... u1e
slides at one time. The subjects were seated at tables before a large rear-
projection photographic screen. They were supplied with answer sheets and
were given taped instructions and the example slides. The 11 pairs of sym-
bol slides were then shown by means of a 35-mm slide projector. After the
three-second exposure of the matrix symbol slide and five-second exposure
of the painted symobls slide, the subjects were given adequate tine (approx-
imately 30 seconds) to indicate their choice by circling the appropriate
letter on the answer sheet.
The same basic administration was used on the color substitution sy~bols,
except that they were conducted in theater-type roans. A total of 85 subjects
viewed those symbols. The instructions given and answer sheets used for both
studies are included in Appendix D.
Data Reduction
The data reduction for this study consisted of a frequency count of the
number of subjects who chose a particular painted symbol as the one most
82
B
FIGURE 25 EXAMPLE SYMBOLS
83
c D
closely resembling the corresponding matrix symbol. These counts were then
converted to percentages of total responses for each pair of slides.
Dsta Analysis ---·-------~--
Subjects were generally consistent in associating one painted symbol
with the matrix symbol. The painted symbols chosen usually appeared to have
the closest structural resemblance tO the matr1x symbol. However, in instances
where there were ambiguities in the matrix symbol, diversion of opinion
occurred. The following findings were drawn from the research.
Group I (Figure 26) -- Responses to the four figures which originally
contained smooth curved lines indicated that the subjects tended to select
the literal interpretation of the matrix symbol more often than any other.
This tendency is shown in the responses to slide pairs I-1, I-3, and I-4.
The exception to that general tendency was the response to slide pair I-2.
There was a considerable division in response to that slide pair.
divis~on could have been due to the fact that three alternatives were feasible
perceptually depending upon the viewer's frame of reference. Thus it apoears
that the use of curved symbols should be avoided in favor of angular, Pore
literal replicas of matrix symbols.
Grouo II (Figure 27t -- Responses to both of these symbols indicated that
the subjects could discern readily the shape of the outline. In fact, many
were able to make the fine discrimination between the different proportions
of the same ~hape figures.
group III (Figure 28) -- Analyses of these data indicated that the sub-
jects readily recognized these unusual and rather complex symbols.
84
Symbol 1..---1 _ ___,
0 0 B 28%
Svmbol I-2 ----'--...,
A B 20% 40%
Symbol I-3
CJO A B 0% 10%
Symbol I-4
A 8 2% 4g_
J
c 0%
c 40%
90%
88°a
D 72%
D 0%
D
D 6%
l
]jPe:-cent of subjects that associated painted symbol with matrix S)'1Ttbol.
FI3UR: 26 GROUP I SYMBOL SELECTION RESULTS
85
B c D 2% 30% 0%
Symbol II-2
A B c D 12% 4% 0% 84%
ll _J Percent of subjects that associated painted sy1ncol with rnatri_'= synbol.
FIGURE 27
GROUP II SYMBOL SELECTION RESULTS
86
A o%Y
A 2%
8 98%
B 92%
c D 2% 0%
c D 6% 0%
!/Percent of subjects that associated painted symbol hith matrix S'.--:-::bol.
GROUP III SY>E30L S~LECTIQ~l RESULTS
87
Group IV (Figure 29) -- Responses to slide pairs IV-1 and IV-3 indicate
that the subjects were awar'? of line \'/eight or stoke width in the formation
of the matrix symbol. This awareness is evidenced by the highest percentage
of responses to the painted symbols being those of bold outlines slightly
greater than the stroke \'lidth to symbol vlidth ratio of the matrix symbol
(Alternatives IV-lA and IV-30). However the painted slides which formed only
an outline (Alternatives IV-lB and IV-3B) also received numerous responses
while those too bold wer~ rejected. These data indicate that in outline
formations it is better to use a painted sign stroke width equal to or
slightly bolder than that of the matrix sign.
Responses to slide pair IV-2 indicate that the subjects had little dif
ficulty associating a solid matrix symbol with its inverse on a painted sign
(Alternative IV-28). The second highest response rate to the solid matrix
block was Alternative 1V-2C, indicating that the positive of the symbol was
perceived by some. These data indicate, and other data from throughout the
study lend support, subjects interpret 1:1hite (on-dark) in matrix lights as
the figure or outline, but when viewed on painted signs, dark (on-white)
is the more ty~ical outline rather than the more literal white-on-dark.
The final slide pair in Group IV, the color substitution slide (Fig~re
30), was studied separately. The objective of this particular segment was to
determine whether a green outlined figure would be selected more often than
an equally correct black outlined form. The percentages shown for slide pair
IV-4 indicate that any effect due to color was not substantial. The black-on
white triangle still received the highest response (57 percent). The two
open triangles received a considerably higher number of responses than did
the solid tria~gles. Thus from this limited study it appears that shape and
88
,-·----------------------------------------------------------~ l l Syr.1bo l IV -1 _______ .__,
DD A 56~
A 2%
A 4%
8 32%
B 82%
SyrnC·ol I\~-3
B 40%
c D 6% 6%
D c D
14% 2%
c D 4% 52%
J-}Pcrcent of subjects that associated painted symbol with matrix s;Fbol.
FIGU~::: 29
GROUP IV SYMBOL SELECTION RESULTS
89
lJ GOW = Green on 1vbi te
BOW = Black on 11/bi te
BOW
57%
BOW BOW
11% 10%
.Y Percent of subjects that assoC:iated painted symbol with matrix s:TnboL
FIGURE 30 . GROUP IV SYMBOLS (CONT.) - SELECTION RESULTS
90
blocking of the symbols are more important than the color.
Summar:Y____Qf_ Results_
1. A group of subjects were tested to evaluate their capability to
associate a symbol on a lamp matrix sign with symbols on a painted sign.
The study results shm,1ed that the subjects more often associated the matrix
symbol with the literal interpretation (exact replica) of the painted symbol,
rather than with rounded figures which could not be constructed with matrix
lights.
For example, a circle, vJhen formed on a lamp matrix, looks more 1ike an
octagon. The subjects saw an octagon instead of a circle which might have
been intended. These results indicate that subjects are aware of the limita
tions of a matrix sign and recognize it is displaying a series of straight
lines rather than arcs. However, these results should not be construed to
mean that they \•JOul d reject the matrix pattern as an acceptab 1 e syr:ibo 1 for
a circular figure such as a zero. In fact, we know that in scoreboards
and other matrix signs octagon-shaped light patterns are perceived to be
zeros.
2. In the study of other sy:r:bols, it 1,;as found that subjects \·;ere co:-i
scious of line widths and shape proportions. These results indicate th3t
shape proportions and line widths should be 2s sinilar as possible to :~2
same proportions and widths of corresponding ~atrix symbols.
3. It was found that the subjects ~ore often associated the ~at~ix
inaae with the inverse of that imaae on a painted sign. A white-on-black
matrix symbol should be a black-on-white Dainted symbol.
91
4. Finally, in a limited follow-up study, it was found that the use
of a color had little apparent effect on symbol selection. The results would
indicate that the shape and blocking of symbols appear to be more critical
than color.
5. In addition to the research findings, a major contribution of this
study was the discovery that many common figures which are easily depicted
on painted signs simply cannot be generated on a 7 X 7 matrix sign. There
fore, in selecting a symbol for a trailblazer, the designer should be aware
that he is either not going to be able to give the code on a CMS or he must
select a very simple symbol such as a square, triangle, diamond, circle or
semi-circle which can be displayed. Also, the symbol should be one for
which there is a common shape name which can be verbalized and possibly be
refered to over the radio. The Group III symbols (the number symbol # and
the snowflake* or asterisk) would probably be more difficult to verba1ize
than the more common shapes and, hence, should be avoided in trailblazing
codes.
6. In summary, the present study did reveal several interesting
findings regarding how people translate figures formed by matrix lights into
painted figures. It was found that they could easily select from several
alternatives the one which rnost accurately matched the pattern of lights
actually displayed. Hm·1ever, these findings should not be construed to r:'ear.
that a less accurate reproduction, when seen alone or in a different context,
would be unacceptable.
7. Future research should address the question of symbol i nterpreta
tion. For example, given a matrix symbol, subjects could be asked to tell
v1ha t the symbo 1 was to see if they perceive it as a common shape. Then the
92
subjects given could be a painted symbol and asked for a similar interpreta
tion. Such a study could provide insight into the transferability of
trailblazer symbols from the freeway (matrix signs) to the surface street
system (painted signs).
93
STUDY SUMMARY
The use of electronic matrix signs as transmitters of dynamic motorist
information messages is a relatively new concept dealing with a wide range
of cor::plex ·issues. Very fev1 design and operational guidelines are avail
able to promote the effective utilization of their capabilities. This
phase of the project research effort was undertaken to provide needed in
formation regarding human factors performance characteristics as related
to various matrix sign design and operational variables. Eight specific
study objectives were initially identified to direct the research effort
toward satisfying the overall goal.
A discussion of the conclusions drawn from the results of this study
with regard to the eight study objectives follows. In a subsequent section,
specific design recommendations are provided.
Any conclusions drawn between sequential and run-on matrix message
displays are drawn with the reservation that all laboratory test data 1·1ere
collected from one sign whose conditions were probably not optimal with
regards to run-on message systems. In addition, all conclusions regarding
run-on signs relate only to one-line signs having the same general display
dimensions.
Conclusions
1. Seqv..enH-:d one-line si;;ns ~;yioba"!JZy offeY' a sr:alZ. advanta'(;e over
one.-:.·t.r:e ?Im-on signs foi' pY'esenting four-word messages to motor-
ist. One Tine sequential or ru:a-on signs s.'J.:Duld not be used to
diaplo..~l 2ight-Llord messaJes to '!!cto::rists.
For messages presented from the first of the Message sequence,
performance was superior 1·1ith sequenti a 1 displays than \'Ii th run-on
94
displays in 7 out of 9 cases. For all four-word message displays,
including beginning in the middle of the sequence, sequential
displays performed better in almost all cases (11/12). For eight-
word messages beginning in the middle of the sentence (See Appendix
A, p. 127), the run-on sign (R-8-2) was superior to the equivalent
sequential sign (S-8-2-2) in all three comparative cases. The
"target value" of the two sign designs was not investigated. Also
not studied were the relative effects of bulb loss (where run-on
displays might offer an advantage) nor system operational co~plex-
ity or reliability.
2. Message length is a very impoPta:nt factor in determining ::-:--:e-:;-:-:er
the complete message can be ~0e,c;roduced. Fcv..I'-7..JCY':i messa;:-?.E' ,-;:::·~~e::I>
. - . -2 ~'._,;~:_~-:8~~
Field test data presented in Figure 10 and also in Table 5 sup-
port these conclusions. Most four-word messages can be successfully
transmitted by the better sign displays at display rates ~f 2.5
seconds per word or slower; whereas, eight-word messages wi:h route
numbers probably cannot be transmitted and successfully reproduced
by more than half of the unfanilic.r free11ay motorists at equ~\Ja1ent
display rates of 1 .0 seconds per word or faster.
3.
95
to four-word message displays &s reduced to a greater extent at
the fasteJ:> display I'ates.
In general, the smaller the time available per word to read
the message, the lower the resulting subject performance in re
producing messages. It is speculated that a critical word display
rate (for a given performance level) may exist for each message
length.
4. No firm conclusion can be drawn as to whether the design variable
11 lines per display 11 is more beneficial than the variable 1\·mrds per
l i ne 11• However, when only two \'JOrds appear in a single message
exposure, a slight increase in performance can be expected if the
second word appears below the first rather than both appearing on
the same line.
5. The study data str>or~gly s~0port the conclusi,.on that messc..g;;.2
sh.m!ld be presented in compact "chunks" at so?:Ie optirrr.;Jn ccr:-.b:>t'°--:;ion
message exposure~ -~ ' • • b n an,u r;1J::e ~ 1,. e. ~ nurn er o J exposv..f'es &:>I t~e
A downward concave relationship in sign effectiveness was found
to exist as the number of lines of a sequential matrix display
increased from one to four.
When four words appear in a message, performance is better if
they appear in one exposure ~vith hm words on the top line and two
words on the bottom. When eight words appear in a message, per
formance is better if they are divided into two exposures of four
words each with a brief time delay (0.7 to 1.0 seconds) separating
96
the two exposures to indicate a separate message or sentence.
These study findings and conclusions are summarized numerically
in the following table.
,----------: Ootimal Si fo (' Given Messa ! -~ ! I I Message Number of Number of Number of ~ford Length Lines/Exposure ~lords/Line Exposures
2 2 l l
4 2 2 l
8 2 2 2
6. In general, messages v1ere reproduced by subjects at about a 10
percent 1 ower level of performance v1hen messages were viewed
starting in the middle of the message rather than at the first
for a given total display time. This finding supports the use of
fewer exposures, each presenting a separate message, so that begin-
ning to read in the middle ~ill not result in a loss of meaning.
For a given level of per-
cent correct subject response (35~, 35~), the percentage of bulbs
out may drop to the level snown as follows:
Motorist Typ~-
Unfamil ·i a r
·Average
Familiar
Maximum Allowable 0 ercent Bulb Loss
Percent Correct Response Criteria 95% Correct 85% Correct
8'1 ·' 18%
1 A o/ '+ ,, 28~~
28~ 44%
97
-
Thus, if the driver may be presumed to be one who has not previously
seen the message, bulbs should be replaced when the bulb loss
reaches 8 percent for 95 percent correct reading.
At the 85th performance criterion for all types of motorist
familiarity conditions, bulb replacement seems to be controlled by
appearance rather than by rea.dabi lity. The researchers feel that
the appearance (and acceptance) degrades to an unacceptable level
at a lower level of bulb loss than that required for reading if
one employs the 85th percentile criterion. A national sign manu-
facturer recommends replacement at 10 percent bulb loss based on
appearance. In designing for the average driver, a maximum of 15
percent bulb loss is recommended.
8. Exceptionally high correct response rates to aU matrix sy~boZs
tested was acceptabZe. In general, symbols containing smooth curves,
and symbols resembling common highway symbols should be avoided in
favor of angular symbols. The painted symbol which will stand for
a matrix symbol should be of similar proportions and line width as
the corresponding matrix symbol and may be a dark outline on a
light background. Color does not appear to be critical to this
identification process.
Although subjects were able to match painted symbols to matrix
symbols with considerable accuracy, this finding should not be
interpreted as meaning that subjects will necessarily interpret
the straight-line symbol as such. Additional research is necessary
to establish how dissimilar a figure must be before it is reject-
ed as a suitable substitute. 98
Design Recommendations
Results of this study suggest several factors which should be taken into
consideration in the design of a dynamic motorist information system. These
rec:orr:mendations should be incorporated into the complete design process to
permit tailoring to fit the specific needs of each system.
Basically, a two-line display is adequate for most applications of elect
tronic matrix signs for providing dynamic motorist information. The length
of a particular matrix sign display \vill depend on the vocabulary and message
font chosen for the system. Thirteen characters oer line appears to ce
about the optimum number with sixteen an upper bound.
Human factors design criteria for designing and operating electronic
matrix signing systems are presented for t\-10 types of design drivers: i)
familiar, and 2) unfamiliar. Another interpretation given to these ~river
types with respect to design conditions are: l) minimum and 2) desirable.
These criteria are presented in Table 10. The display rate is tr,e raxinu:n
time permitted to read each word of the message being statically displayed
from the moment the design dri'1er first comes within his legibility distance
of the sign (as defined by the product of the letter height of the characters
and the design legibility for the cfri'ler) u;itil the "11ull distance'' fr08 the
sign is reached (arbitrarily defined by a maximum permitted reading ang1e to
the sign of 10 and 5 degrees for f2miliar and unfam~1iar drivers).
The required character heights for the conditions defined varies wi~h
the 85th percentile speed of the traffic approaching the sign. The minimum
1 etter heights · .. 1hi ch satisfy the pre vi ousl y noted human factors requirements
are presented in Table 11.
99
Table 10. Selected Human Factors Requirements Design Criteria for Matrix Signs
r-oes tgn Criteria
l. Display Rate (sec./word)
a. Four-word messages
b. Six-word messages
c. Eight-word messages
2. Legibility (feet/inch)
3. Null Distance c (feet)
a. Overhead sign
b. Roadside sign
Familiar Driver
0.50
0.75a
1. ooa
42
100
175
a Estimated.
b Message must be repeated at sign or downstream. c Distance immediately upstream of sign ineffectiv
Subtract this distance from overall legibility d obtain total available viewing dist2ncc and read traffic speed.
100
Unfamiliar Driver
0.75
l .OOa
l.OOb
35
200
350
for reading sign. stance provided to . . ,... . ng :1ne ror a give~
-
Locc.tion (' .c ) I
-~-~
Overhead
Roadside
Table 11. Recommended Minimum Letter Heights for •Various Message Lengths, Sign Locations and Approach Speeds for Familiar and Unfamiliar Driversa
85 %-tile Message Letter Heigbts (Iocbes) Approach Length Familiar Unfamiliar
Speed (MPH) (\'lords) Driver Driver
40 4 5 11
6 9 16
8 14 33
50 4 6 12
6 10 18
8 16 39
60 4 7 13
6 12 21
8 19 46
40 4 7 15
6 11 20
8 16 37
50 4 8 16
6 12 23
8 18 44
60 4 '"' 18 ::;
6 14 25
8 21 50
aEight-\,1ord messages cannot be recommended for complex messages based on study results. In addition, use of letter heights less than 10 inches should be avoided for freeway signing unless experience justifies other-wise.
101
..
! I
i I I
l I
I I I I
I I
I i
I I
I I I
I
I i I I I I
I '
As noted previously, two-line signs appear to be highly cost-effective
compared to other candidate configurations. If all of the words in a message
:.· ~ 1 : -.!.' "j +· >'{I' I I I ..... on the two-line sign, the entire message should be displayed in one
exposJre (S-4-4-2). If the entire message will not fit on two lines it should
be c!isplayed in two exposures. Two long words, or the equivalent, may be
displayed per exposure with one word on each line (S-4-2-1). For one-line
signs, it is almost equally effective to display two words on one line and,
in the next exposure, the other two words on a line (S-4-2-2). If economic
constraints necessitate the use of a one-line sign with few characters per
exposure, either a sequencing mode or a moving mode may be used (S-4-1-1 or
R-4-2). However, this research favors the use of the sequencing mode.
All non-static messages should be followed by a brief 11 blank-out 1' mes-
sage to indicate the end of the message. Blank-out messages should be 0.7
to 1 .0 seconds in length.
Bulb loss on electronic matrix signs should not exceed 15 percent based
on legibility criteria. This bulb loss rate is such that replacerr:e::t due to
poor appearance may govern the replacement program since the credibility of
the sign may be questioned by motorists viev1ing a sign having more than 10
percent lamp outages.
The use of matrix sign symbology to introduce a route trailblazer sym-
bol, \'thich \'!ill be presented on painted signs along the des-lgnated roc;te.
can be expected to perform successfully if a few guidelines are followed.
The basic requirement suggested by the data is that trailblazer symbols
should be wad2 of short, straight lines rather than smooth curves. Like-
wise, symbols resembling common highway symbols should not be used. The sym-
bo1s shauld be capable of being generated on the matrix sign and should be
cc-:-:r"c'; shapes for which a name can be assigned. 102
REFERENCES
1. Dudek, C. L. Human Factors Requirements for Real-Time Motorist Information Displays. 11 Vol. 2, State-of-the-Art Real-Time Motorist Information Displays, Texas Transportation Institute, February 1978.
2. U. S. Statistical Abstract, U. S. Government Printing Office, Washington, D. C., 1971.
3. Pierce, J. R. 11 Symbols, Signals and Noise." Harper and Brothers, New York, 1961, p. 236.
4. Mitchel 1, A. and Forbes, T. W. '1 Design of Sign Letter Sizes. 11 ASCE Proceedings, No. 67, 1942.
5. Miller, G. 11 The Magical Number Seven, Plus or Minus Two." Psycho1ogical Review, Vol. 2, 63, 1956, pp. 81-97.
6. !ES ~ighting Handbook, Fourth Edition, Chapter 16 --"Lighting for .Advertising." Edited by John E. Kaufman. Illuminating Engineering Society, New York. 1968.
7. Bogdanoff, M.A. and R. P. Thompson. 11 Evaluation of ~.Jarning and Information Systems, Part I, Changeable Message Signs. 11 Freeway Operations Branch Report No. 75-5. Low Angeles Area Freeway Surveillance and Control Project. September 1975.
8. Helgerud, Leif. tion Institute.
"A Daytime Study of Driver Legibi1ity." Texas TransportaUnpublished report.
9. 11 Binocular Visual Acuity of ,!1,dults, 11 'lationa1 Center for Health Sta~istics, Series II, Number 3, U. S. Department of Health, Education and Welfare, June 1964.
10. Telephone discussion with representa:1ves of American Sign and Indicator Corporation, Spokane, ~·Jashington, A.pril 18, 1975. Project Technical Memorandum 75-37.
11. Van Coff, H. P. and Kinkade, R. G. 11 Human Engineering Guide to Equiprrent Design. 11 r.~csra~·1-Hill, 1?72.
103
APPENDIX A
EXPERH1ENTAL DESIGN
FOR
LABORATORY EVALUATION
OF
MP.TR IX DI SPLA.YS
104
EXPERIMENTAL DESIGN
Flashing vs. !1oving Messages
OBJECTIVES
• To determine the relative effectiveness of flash vs. moving
messages
•To determine the relative effectiveness of one-line vs. two-line
vs. four-line presentation for four-word and eight-word messages
FACILITY
Media Lab
TEST EQUIPMENT AND INSTRUMENTATION
1 Film of the Changeable Message Sign
fl Inst ru c: ti on s
t Projector
@ Tables and Chairs
SUBJECTS (Total Number 200
Categories:
11'.\ge
Sex
Education
TEST PERSONnEL. ~.:rn SUPPORT
One admin~strator required
103
TEST SCHEDULE
1-5 subjects per session until 200 subjects obtained.
TEST DESIGN
Independent Variables
• Message Length
1 Type Presentation
• Presentation Time
• Initial Message Start
1 Words Per Line
Criterion Variables
1 Percent Correct Responses
Controlled Conditions
• Blank to end messages
•No time between flashes except at end of message
1 5 lamp columns between words on run-on
• 200 subjects
1 2 Message lengths; 8 and 4 words
a 2 Message start points; First, Middle
o 3 Presentation times per message length
106
Statistical Design (Cont.)
4 l·Jord
c::: 1.0 second (4 word/sec)
c = 2.0 second ( 2 word/sec)
c = 3.5 second (1. 14 \\lord/sec)
c = 2.0 second ( 4 \·1ord/sec)
c = 3.5 second (2.28 \'lord/sec)
c = 7.0 second (1.14 v1ord/sec)
f1iddie
4 \ford
c = 2.0 second ( 3. 75 v1ord/sec)
c = 3.5 second (2.14 word/sec)
c = 7.0 second ( 1 n-, '• ~a'1'-o ) \ .v1 ~1oj ·. ~.._c
e 4 Presentation modes; l , 2, 4 line, run-on
• 2 Words per line; 1, 2
• 2 Randomized scripts (replications)
TEST DATA ANALYSIS
Data Reduction Methods
c = 3.5 second ( 4. 28 1t1ord/sec)
c = 7.0 second (2. 14 word/sec)
c = 14.0 second (1.07 ~"ord/sec)
Prepare curves representing percent correct response as measure
of effectiveness of prEsentation mode.
Tes ts of Si qn ifi c_a_r1 ce_
Differences in mean values or oercent correct response.
ATTACHMENTS
• Presentation Mode and Rate
a Detail Procedure and Instructions
l'i Data Sheets
107
PRESENTATION MODE AND RATE
M::SSAGE ORDER (Beginning at First of Message)
Cycle Time (Seconds)
Desi 9!2- c = 3.5 c = 2.0 c = 1. 0 -
S-4-1-1 I lfordl .88 0.5 0.25
[_!:lord I .88 0.5 0.25
r v1ord I .88 0.5 0.25
[t-iord l .88 0.5 0.25
S-4-2-1 iford 1. 75 l. 0 0.5 \·Jo rd
\ford 1. 75 1.0 0.5 1.-Jord
S-4-4-1 \~ord 3.5 2.0 l. 0 \ford \!Jo rd
LJ:!__ord
S-4-2-2 I Word \lfo rd 1. 75 l. 0 0.5
Li'""'d --;-i
l. 75 l. 0 0.5 ', 1_ I \!lord J
S-4-4-2 btord \!Jard j 3.5 2.0 ; . 0 ord \ford I
R-4- l I iforcfl ' .88 [) r l . :) 0.25
I iford I y .88 0.5 0.25
r\.Jor~ y .88 0.5 0.25
i i.Jord ! ~__J
y .88 0.5 0.25
108
Cycle Time (Seconds)
Q.e? i__gi) __ c = 3.5 c = 2.0 c = 1.0
t·~ - 4-~ 2 h09r~ ' l. 75 1. 0 O.S
LJJ9_r:_d_
f-Jf?!Ij v , ,,- -; _:~Jrd i • I ~-J l. 0 0.5
c = 7.0 c = 3.5 c = 2.0 S-3-2-2 c-·------ J _:.:lord _}ford l. 75 .88 0.5
[i~ord ~Jo rd I l. 75 .88 0 i:; -~
[~Jard Word I l. 75 .88 ·1 ,... L!.)
[]ord Word I 1. 75 .88 0.5
S-8-4-2 lfo rd vJord 3.5 1. 75 1. 0 \ford l'1ord
~ford \;lord 3.5 l. 75 1. 0 Word Word
S-8-8-2 \ford \ford 7.0 3.5 2.0 ,---,--, d \ford ~~_!::_:._ i ',iord \ford I. l'.2 r-c[ \·Jo rd
' R-8- l [Hord] . 88 . 44 0.25
LJford l ' .88 .44 0.25 r--;-1--d---,
' .38 4Ll 0 ''C: i v.or i . 'J .. C....J
U~g-~ v .38 .44 0.25 .------:Jl y . 88 .44 (\ ?r::: LJj_:2_!_~J 'J • '-....,:
[\.Jord]
' .88 .44 0.25
l___\11 o r.9-_] y .88 .44 0.25
r Wore[] y .88 .44 0.25 ...
109
Cycle Time (Seconds)
Desi g~- c = 7.0 c = 3.5 c = 2.0
f~-8-2 [;-rdl y 1. 75 .88 0.5 ord I
11:Jord] tt~:-g):"(f . y l. 75 .88 0.5
tford y l. 75 .88 0.5 Word
\'Jo rd y 1. 75 .88 0.5 l·lord
MESSAGE ORDER (Beginning at Middle of Message)
c = 7.0* c = 3.5* c = 2.0*
S-4- 1- l [Word] . 93 .47 .27
OTuiiU . 93 .47 .27
UUank] l.40 .70 .40
I ~lord I . 93 .47 .27
! \''oi:::.QJ . 93 .47 .27
: 1-10-rci' i.____:._:_____~·----' .93 .47 .27
[1"iorG1 .93 .47 .27
S-4-2-1 ~lord 1.86 .93 5 ·1 • "T
I.ford
I Bl an k I 1.40 . 70 .40
TWord I J Word I
1.86 .93 .54
i
~Jo rd !
1.86 .93 .54 l Word
110
Design_
S-4-4-1
S-4-2-2
S-8-4-2
fl:iordl ~----- -~ : \~ord ; f-+----~ ! i.Jo rd l : ~lord i ------- ·--___;.
r Blank I
I \.Jord I Word iford
! \forci1 L.: __ J
\ iford 1.Jorct I
[Blank[ .---
1..io rd I l lford
J Word ~lord l
r-~fo ref-· 1Jord ~ iWo-r:;r--~19_ r (fj
i,.'; Q rCJ '.~ (Yt"" r-j ~
~ \·Io r~g -_ i~~:ij
c = 7.0*
1.86
l.40
3.72
1. 86
1.40
1.86
1 .. 86
c = 14. 0*
3. 72
2.80
3. 72
3.72
Cvcle Time (Seconds)
c = 3.5* c = 2 ·"* - . ·.)
. 93 .54
.70 .40
1.86 1. 08
.93 .54
.70 .40
.93 .54
.93 .54
c = 7.0* c = 3.5*
i.86 .93
1.40 .70
1.86 a~
1. 85
* The initial study used 20% of the cv:le time blank at the end of the seauencing mode and an end chdracter of 9 lamp columns to designate the end of a run-on r1ode. The suoolementary studv used 0.8 seconds blank time at the end of the sequencing :~1ode.
111
Design
S-8-8-2
R-8-1
Word Word I \.lo rd_ lfo rd ! \ford vlord , ___ _ WP rtj l~o rd
I Blank I Vlord Word Word vJord vlord ~ford
I ~ford vlord
! vJord I I 1.Jord I I vlord I
I tford I
[sJillaj I viord J
[Jford]
I t.fo rd ;
! \ford : ___ J
!_\-Jo rQ.._
i \·J_Q_rd :
I \fordJ
iGJord . ~-
y y y y
v y y v y y
' y
c = 14.0*
3. 72
2 .. 80
7.44
l. 08
1. 08
1.08
1.08
.90
1. 03
l. 08
1. 08
1. 08
l. 08
l . 0(3
l. 08
1. 08
112
Cycle Time (Seconds)
c = 7. O*
1.86
1.40
3.72
.54
.. 54
.54
.54
.44
.54
.54
.54
5,c.
5.'1 • "T
.54
.54
5'1 . ,
c = 3.5*
. 93
.70
• 93
"'.l 1 • ._, I
. 31
. 31
. 31
.22
. 31
..., 1 •.) I
.31
'), .J !
..., , ~ .j !
~; ... ._; ~
"'.l' ·"I
.31
Design_
R--8-2 µi_g_rd ! y L v!o_rd_J
~~lord \ I \ford j ~ [Blank!
vlord I y Word
I Word I I Word I ' ! \>Jord I y r I Word I
I \ford ! y i-u--~
L11oi:Q_j
I Blank I
R-4·- l [}ford--1 T
[ \.JoI_c!_] l B 1 an k
I \.lord I f I Word I y
[}To-r?] ' L__\.iord I y
c = 14.0*
2. 16
2. 16
.90
2. 16
2. 16
2. 16
2. 16
c = 7.0*
t /" .'l I ~: .. :
3. 72
1. 08
44
1.08
l. 08
1. 08
1 no ~ .. vu
113
Cycle Time (Seconds)
c = 7.0*
l.08
l .08
A4
l.08
1. 08
l.08
l.08
c - 3.5*
0" . _,.)
.70
l. 86
~/l . ._,-..
.22
. 54
.54
.54
.54
c = 3.5*
.62
.62
.22
.62
.62
.62
.62
c = 2.0*
.54
. 40
1. 03
.31
. .31
. 12
. 31
.31
. 31
. 31
Desi .9._t]_ Cycle Time (Seconds)
c = 7.0* c = 3 ~* • ::i c = 2.0*
R-4-2 Word y 2. 16 1. 08 .62 \ford
I Blankj .44 . 24 .12 -
vJord y 2. 16 1. 08 .62 Word
lford l 2. 16 1.08 .62 Word
Cycle Time (Seconds)
Desi 9.D_ c = 14.0* c = 7.0* c = ".) '""* '-'• ::J
S-8-2-2 lford \>lord I l. 86 .93 .47
vlord i..Jo rd I 1. 86 .93 .47
I Blank I 2. 80 1. 40 . 70
lford ifo rd I 1. 86 .93 .47
[ \ford t.Jord I 1 . 86 . 93 .t17
I Hord v1ord I 1. 86 .93 .47
11fnrd \·!or:d I l.86 .93 .0
114
ORDER A
FIRST faJlD MIDDLE MESSAGE START -- AF, AM)
Presentation Mode
*
S-4-4-1 CAF* = 3.5 CAM*= 7.0
R-4-2 CAF 2.0 CAM = 3.5
S-8-8-2 CAF = 7.0 CAM = 14. 0
S-;3-4-2 CAF = 2.0 CAM = 3. 5
S-4-2-1 CAF = l .0 Cl\M = 2 "0
R-8- l CAF = 7.0 CAM = 14. 0
C.AF = Cvcle C.A}'ri Cycle
t t
me in r;ie in
SLIDE AREA AHEAD NEXT
IROAD ALERT I AUTO RADIO
SHARP TURN AHEAD NEX: US-69 EAST SLOH 00\,JN
~~~~~--~-_j
r-· ----------1 SLICK RO,~D
U ~ -38 \JCC:T -i ....; . :d :...._.....; j
i USE CHA LtS '. L----------------------
Ul.NE I BLOCK I JUST l A.YE.!'\D
I H~IY-9 I EAST EX IT ! RIGHT l seconds for message order "A" beginninq at first of message. seconds for rness2,~e order "A" beginning at widdle of r.iessage.
115
Presentation Mode
S-4-2-2 CAF = 3.5 CAM= 7.0
R-8-2 CAF = 7.0 CAM = 14 .. 0
S-8-2-2 CAr = 2.0 CAM = 3.5
S-4-1-1 CAF = 3.5 CAM= 7.0
R-4--1 CAF = 1.0 CA~~ = 2. 0
S-4-4-2 CAF = 2.0 CAM = 3.5
R-8-2 CAF = 3. 5 CAM = 7. 0
116
Message
I TOLL STOPS I NEXT RIGHT I .
I SNOW SL IDE I AHEAD NEXT I I I-835 \•JEST I BEST ROUTE I
ROCK SLIDE
ALONG TURN
US-23 EAST
MERGE LEFT
REST
AREA
ALm!G
ROAD
I AUTO I WRECK i SLOW I DOWN!
SHARP TURN NEXT RIDGE
I ROAD \·IORK I ALONG TURN!
jUS-81 WEST I MOVE RIGH~
Presentation Mode
S-8-4-2 CJ\F = 3. 5 C,£1.M c-= 7. 0
5.;.4 .. 2-2 CAF = 2.0 CAM = 3. ~i
R-4-1 CAF = 2.0 CAM = 3.5
S-8-2-2 CAF = 3.5 CAi'v1 = 7. 0
S-4-4-1 CAF = l .0 CA>: " 2. 0
R-4-2 cr~F = 3. 5 CAM = 7.0
S-4-2-1 CAF = 2. 0 CAM = 3.5
117
Message
!Slo~·I-TRU~ ~NG R0,1\D I
US-81 WEST USE BYPA?_S __
TRUCK TOLL
ROUTE NEAR
!sLIDE I AREA !ALONG I ROAol
ROAD ALERT
!AUTO RADIO
H;.;y -6 \·iE:ST
r ST'\Y TU'~'c 1 ! ._L __ ._r. __ ~--'
REST 11.REA AHEi\D i·lEXT
SHARP TURN
NEXT AHEAD
Presentation Mode --·
R-8-· l CAF = 3.5 Cft.M 0: 7. 0
S-4-1-1 CAF = 2.0 CAM = 3.5
R-4-1 CAF = 3.5 CAM= 7.0
S-8-8-2 CAF = 2.0 CAM = 3. 5
S-4-4-2 CAF = 3.5 CAM = 7 .0
S-4-4-2 C.i\F = 2. 0 CAM = 3.5
CAF = 3 .. 5 C.t\.M = 7. 0
118
Message
lsLm~ I TRUCK I ALONG I ROAD!
I I -415 I WEST I MERGE I LEFTI
SLICK
ROAD
AHEAD
NEXT·
l RO,i\D I ALERT I A.HEAD ! NEXT!
TOLL STOPS NEXT AHEAD I-950 EAST MERGE LEFT
ROCK SLIDE NEXT RIDGE
SLIDE AREA ALONG TUR~J
SHARP TURN NEXT RIDGE HWY-5 EAST SLOW DOl~N
Presentation Mode
S-4-2-2 CA.F = l . 0 CAM = 2.0
R-8-1 CAF = 2.0 CAM = 3.5
S-4-2-1 CAF = 3. 5 CAM= 7.0
R-4-2 CAF = 1.0 CAM = 2.0
S-8-2-2 CAF = 7. 0 c 1'.J.".1 = l 4 . n
S-8-4-2 CAF = 7.0 CN·: = 14, 0
S-4-4-1 CA.F = 2 .0 CA.M = 3 .5
l 'O L
Message
TOLL STOPS
~_N_E_AR ROUTC"J
I SHARP I TURN I NEXT i RIDGE I I us-69 I t-IEST I sLm,1 ! DOWN I
SNOW SLIDE
ALONG TURN
REST AREA
r-------··-··------i
L H\·IY -7 l·IE.S: _ ___!
l T.!\KE BREA~
TP.UCK TCLL .AHC:.AD NEXT
I-835 ~JEST EXIT RIGHT
--, sNm~ 1
SLIDE AHEAD NEXT
Presentation Mode
S-4-1-1 CAF = 1.0 CAM = 2 .. 0
R-8-2 CAF = 2.0 CAM = 3. 5
120
Message
ROAD
ALERT
NEAR
ROUTE
I ROCK I SL IDE I NEAR I ROUTE I ]HIJY-6 j EAST !STILL loPrnl
Presentation Mode
S-El-8-2 CBF* = 7.0 CBM* = 14.0
S-8-4-2 CBF = 2.0 CBM = 3.5
S-4-2-2 CBF = 3.5 CBM = 7.0
R-8-2 CBF CBM =
S-4-4-1 CBF
2.0 ') ,.. J. :J
1.0 2.0
R-4-2 CSF CBM
::: 3. 5 = 7.0
ORDER B
(FIRST AND MIDDLE MESSAGE START -- BF, BM)
Message
ROAD vlORK ALONG ROAD HWY-5 EAST MERGE LEFT
SNOW SLIDE NEXT RIDGE
I-270 EAST BEST ROUTE
SLICK ROAD
ALONG TL;R~~
!AuTo I i~RECK ! JUST I !l.HEr:.o I I I-415 I \!EST ! SL.G;·I I DOW~\
ROCK SLIDE
ROUTE
!LANE I BLOCK I AHEAD l NEXT I
* BF = Q\i -'-"• \ -
Cycle Cycle
t "'e n seconds for message order "B" beginning at first of message. t me n seconds for message order "B" beginning at middle of message.
121
Presentation Mode
S-8-·2-2 CBF = 7.0 C:3M = 14. 0
S-4-1-1 CBF = 1.0 CBM = 2.0
S-4-4-2 CBF = 3.5 CBM = 7.0
R-4-1 CBF = 2.0 CBf.l = 3.5
S-4-2-1 CBF = 1.0 CBM = 2.0
R-8-1 CBF = 7.0 CBM = 14.0
S-4-4-2 CBF = l.O CBM = 2.0
122
Message
TOLL ROUTE
AHEAD NEXT
US-38 WEST
ENTER HERE
SLIDE
AREA
NEXT·
RIGHT
TRUCK TOLL ALONG RO/.\D
I ROIK I suor I ~JFAD l Rt'UTEi 1...... I c I··- 11\ I .. u i
TOLL STOPS
NEXT AHEAD
lsLow jTRUCK !ALONG I ROAD!
r us-69 1 EAST 1 MERGE 1LEFT1
SHARP TURN AHEAD NEXT
-·
Presentation Mode
S-8-8-2 CBF = 3.5 CBM = 7.0
S-4-1-1 CBF = 2.0 CBM = 3.5
R-4-1 CBF = 1 .0 CBM = 2 .. 0
S-8-4-2 CBF = 3 .. 5 CBM = 7.0
S-8-2-2 CBF = 3.5 CBM = 7. C
S-4-2-.2 CBF = 2.0 CBM = 3.5
R-8-2 CBF = 3.5 CBM = 7.0
123
Message
SLICK ROAD ALONG TURN I-835 \~EST USE CHAINS
ROAD
ALERT
AHEAD
NEXT
I 0 0~0 I IJQRK I ~i;-x.,.- i ·1 •·r-•n I j •·\ r', I !. J I' C, i ; :-,r:c.i-li.J I
TRUCK TOLL NEXT RIGHT
H1,.JY-7 ':JEST i
TOLL FP.EE I _____,
~_R_E_ST AREA __j
[ US-23 E.AS:
L TAKE BREAK
r-1 SLICK ROAD
NEXT RIDGE
jSHARP !TURN I NEXT !RIDGEi
!HWY-9 jEAST !SLOW !DOWN I
Presentation Mode
S-4-2-1 CBF = 2.0 CBM = 3.5
R-8-1 CBF = 3.5 CBM = 7 .. 0
S-4-4-1 CBF = 2.0 CBM = 3.5
R-4-2 CBF = 2.0 CBM = 3.5
S-8-2-2 CBF = 2.0 cs111 = 3. 5
S-4-2-2 CSF = l .0 CBM = 2.0
R-8-2 CBF = 7.0 CBM = 14.0
124
Messaqe
LANE BLOCK
JUST AHEAD
!ROAD !ALERT !AUTO IRADIOj
!us-s1 I WEST lsTAY ITUNEol
SLm~ TRUCK ALONG ROAD
!TRUCK I TOLL !AHEAD I NEXT!
TOLL STOPS
NEXT AHEAD=1
1.;s-69 usr~
USE BYP.L\SS
ALONG R0.11.D
!ROAD !SLICK I ALONG I TURN!
I H\·IY-5 I vlEST I EX n I RIGHT I
Presentation Mode
S-4·-1-1 Cl3F = 3.5 C3M = 7.0
R-4-1 CBF = 3.5 CBM = 7.0
S-4-4-2 CBF = 2.0 CBM = 3 .. 5
S-4-2-1 CBF = 3.5 CBM = 7.0
R···8-· l CBF = 2 n CBM = 3.5
S-3-4-2 CBF=7.0 CBM = 14.0
S-4-4-1 CBF = 3.5 C8M = 7.0
125
Message
[ SNO\~ =:J SLIDE I
oLONG
TURN
!REST !AREA INEXT I RIDGEi
TRUCK TOLL SLOW DOWN
AUTO ~~RECK
Jl,LONG RO;~D
!SLIDE !AREA j ALONG l ROAD!
II-270 !WEST I MOVE !RIGHT!
SNO\~ SLIDE
H\·JY-6 WEST NEXT RIGHT
TOLL ROUTE AHEAD NEXT
Presentation Mode
R-4·-2 C!3F l . 0 CBM = 2. 0
S-3-8-2 CBF = 2.0 CBM = 3.5
126
Message
I SLOW I TRUCK I JUST I AHEAD I
ROCK SLIDE -ALONG TURN I-950 EAST STILL OPEN
Manual Start
Auto Sl·ide Advance and Projector on
Auto Slide Advance
Auto Projector Off
Auto Film Projector On
Auto Film Projector Off
Auto Tape Stoo
TEST PROCEDURE AND INSTRUCTIONS
This is a study to determine how fast a driver can correctly
read various messages regarding traffic conditions along a
roadway like this:
(Slide of Freeway)
These messages might be formed and flashed on an electric
display sign similar to the one now shown on the screen, cr.d
could be changed at any time.
(Slide of Matrix Sign)
(Pause)
In this study, we will show 36 different messages having
various message lengths presented in different ways, and for
various lengths of ti~e- The message will be flas~ed sn the
screen, and as soon as it is removed. look at the answer
sheet which your experimenter has passed out. Here, beside
each corresponding message number is a blank soace. w~ite
the message which you read on the screen in the corres~onaing
blank Here is an example ~essage.
(Project Message on Screen)
Just to help you keep track and to be ready, I will announce
each message as it appears on the screen. Now, are there
any questions?
127
Manual Tape Restart
Auto Film Projector Start
Auto Film Projector Stop
Auto Film Projector Start
Auto Film Projector Start
Auto Film Projector Stop
All right, the first message will appear in just a second.
Remember, as soon as the message is flashed on and off the
screen, look at the answer sheet and write the message you
read.
(Pause)
Here is the first message that might be flashed on the sign:
(First Message)
Here is the next message:
(Second Message)
Here is the last message:
(Last Message)
OK, that completes this study. Thank you very much.
Initial Instructions. 3.0 r;d nutes
Questions .. 2.5 mi;;utes
Messages (36) and Reconstruction. .18.0 minutes
Additional Instructions, Closing. 1. 0 Jilinutes
Prograrn~ing Lost Time 0.5 minutes
Total Time ... .25.0 minutes
128
4 Films
2 - Random order of messages
2 - Message beginning points
Subjects: 50 / Block
Message Order 11 A11
Start at Beginning of Message
Message Order 11 8 11
Start at Beginning of Message
~?O
Message Order "A" Start at Middle cf Message
Message Order "B" Start at Middle of Message
ANS\·JER SHEET
ORDER A, B
NAME DATE TIME ---·-------- ------- -------
l. ----------------------------2. ----------------------------3. ___________________________ , 4. ---------------------------5. ---------------------------6. ----------------------------7. ----------------------------8. --------------------------9. ___________________________ ,
l 0. --------------------------11. -----------------
12. --·
13. --14. ----------------------------
15. ---
16.
17.
130
18.
19.
20.
21.
22.~~-~-~--~-~~~~~~~~-----~---
23._~--~-~--~~---~-~~-~-~--~-~~
24.~~~~-----~~~~-~~-~-~-~~~--~~
25.~~----~~~~-~~~~-~-~~~~----~
26.
27·--~~~~~~-~--~~~~-~~--~--~-~~
28.~~--~~~~-~---~---~~--~--~---
29.~-----~--~----~~-~~~-~~-~----
30.~~-~--~--~~-----~--~----~---
31.~~-~------~----------~----~---
32·----------~-------~ ---~--~---~-33. __________ _
34. ---
35. -----·------------.·-~--------- ·-----------------
131
100
90
80
7
SUMMARY OF RESPONSES TO FOUR-AND EIGHT -~JORD MESSAGES-MIDDLE OF MESSAGE PRESENTATION
(S-4-4-2) (S-4-4-1) (S-4-2-1)
S-4-4-2 (S-4-2-2)
S-4-4-1 S-4-2-2
S-4-1-1
S-4-2-1
132
·-
APPENDIX B
EXPERIMENTAL DESIGN
FOR
FIELD EVALUATION
OF
MATRIX SIGNS
133
EXPERIMENTAL DESIGN
l. Title
Dynamic Display Field Studies
a. Deti:~rmine the relationship between laboratory results and subject response in a 11 loaded 11 and an 11 unloaded 11 driving situation. The following specific criteria will be considered:
l. Determine whether subjects can perform as well (or better) in an actual driving environment as in the laboratory.
2. Determine whether the laboratory rank order of the four-word message formats will hold true for the field situation.
3. Determine whether the correct response rate for eight-word messages will improve under more realistic conditions.
4. Determine whether the use of a slower word rate will improve correct response rates for both four and eight-word sessages.
5. Determine ~vhether there is a substantial difference in subject performance when the subjects are required to perform confined lane tracking while reading the sign.
b. Estimate the approximate legibility of the 18-inch lamp matrix sign.
3. Facility
Proving Grounds at Texas A&M Research Annex
4. Test Equipment and Instrumentation
e Test subject car
• Variable message lamp matrix sign
o Traffic cones
e Two-way radios
134
,_
5. Subjects
40 total subjects (20 per condition)
6. Test Personnel and Support
e Variable matrix sign operation
5 Test subject experimenter
7. Test Schedule
One subject per hour session until 20 subjects tested for each of the two conditions.
8. Test Design
8. 1 Independent Variables
a. Message length
b. Display time
c. Words per line
8.2 Dependent Variables
a. Legibility distance
b. Percent co:rect response
l ) Message reconstruction
Reasona.ble maintenance of travel speed ("loaded" condition only)
., \ 3) Displace~ent of lane cones ("loaded" condition on'.
8.3 Controlled Conditions
a. Initial message start
b. Type of presentation
c. No blank time between sequences except for eight-word message at 0.5 seconds/word display rate
d. Two-second blank beh1een four-•t1ord "chunks 11 in ei ght-'.'/Ord message at 0.5 seconds/word display rate
e. Blank to end messages
f. Vehicle speed
8.4 Statistical Design
a. 20 subjects
b. 2 message lengths: 4 and 8 words
c. Display times of 0.5, l .0 and 2.0 seconds/word for all messages
d. 2 variations of words per line: 1 and 2 words
e. 2 randomized message orders
9. Test Data Analysis
9. l Message Displays Data Reduction Methods - Prepare tables representing percent correct response as a measure of effectiveness of presentation mode for both 11 loaded 11 and "unloaded" conditions
9.2 Legibility Distance Data Reduction Methods - Prepare histogra~ representing legibility distance as a function of percent of subjects
10. Attachments
0 Detailed procedure and instructions
136
MESSAGE PRESENTATION
Seq'Jence Number Display Time
2.0 sec/wd l. 0 sec/\·1d 0.5 sec/wd
l . \</ORD \iJORD I 4.0 2.0 l. 0
(S-4-2-2) WORD WORD[ 4.0 2.0 l.O
2. I \i!ORD I 2.0 1. 0 0.5
(S-4-1-1) l l•IORD I 2.0 l. 0 0. 5.
f l·JORD ] 2.0 l. 0 0.5
I WORD I 2.0 l. 0 n -~ '._,.-..I
3. WORD WORD 8.0 4.0 2.0 WORD WORD
(S-4-4-2)
4. WORD 4.0 2.0 1. 0 \·IORO
(S-4-2-1) \·!ORD 4.0 2.0 l. 0 \·!ORD
[J]SJ----· 5. - :- " r. .1,1 \·J0'-'U -~~- I\ I , , , .. ,, "
_·1.:_J~<u \~ORD 8.0 4.0 (S-8-4-2)
\·!ORD 8.0 4.D
6. \·JORD \~ORD \'IORD WORD 2.0
(S-8-4-2) Blank 2.0 \iJORD viORD \~ORD \·!ORD 2.0
1 ~7 .) '
PROCEDURE
l. Subject in test car positioned well back of legibility distance.
2. One of three legibility test words displayed on sign.
3. Subject receives instructions to move forward at a low speed and to call out the test word when he can read it.
4. Administrator records the legibility distance and asks the subject to return to the starting point for remainder of legibility runs.
5. Subject receives additional instructions regarding the message display studies.
6. As subject proceeds in test car, administrator depresses the key on the radio at subject 1 s legibility distance recorded in Step 4.
7. At the sound of the key being depressed, the sign operator activates the particular sequence to be displayed on the run.
8. After the subject passes the sign, the administrator asks the subject to repeat the message. Administrator records those portions of the message which the subject correctly recounts as well as those parts incorrectly recounted.
9. The sign operator blanks the sign and awaits the next run.
10. Subject returns to the starting point for the next run.
11. Steps 6-10 are repeated for each run.
LEGIBILITY TEST WORDS
A) BOAT
B) BOOK
C) ROCK
138
M1 - ROCK SLIDE NEXT RIDGE
M2
- SHARP TURfl NEXT AHEAD
M3
- SHARP TURN NEXT RIDGE
M4 - SNOW SLIDE ALONG TURN
M5 - SLIDE AREA ALONG TURN
M6
- TOLL STOPS NEXT AHEAD
M7 - TOLL ROUTE NEXT EXIT
M8 - SLICK ROAD ALONG TURN
Mg - AUTO WRECK ALONG ROAD
M10 - TRUCK TOLL SLOW DOWN
M11 - ROCK SLIDE NEAR ROUTE
M12
- LANE BLOCK JUST AHEAD
M13
- SLQ!.-J TRUCK ,L\LO':G ROAD
M 1 ~ - TRUCK TOLL NEXT RIGHT ~ HWY-7 WEST TOLL FREE
M15 - ROCK SLIDE Jl.LO:;G ~-Li?::
US-23 EAST MERGE LEFT
~.,
"16 s~ow SLID~ B~ST ROUT~ HWY-6 WEST NEX~ RIGHT
M17 - SLOW TRUCK ALONG ROAQ US-81 WEST USE SYPASS
MESSAGES (Unloaded)
139
M1 - REST AREA ALONG ROAD
M2 - TOLL ROUTE NEXT RIGHT
M3 - SNOW SLIDE AHEAD NEXT
M4 - ROAD WORK NEXT AHEAD
M5 - ROCK SLIDE ALONG TURN
M6 - SLOW TRUCK JUST AHEAD
M7 - LANE BLOCK NEXT EXIT
M8 - AUTO vJRECK MERGE LEF_T
M9 - TOLL STOPS SLOW DOWN
M10 - SLICK ROAD SLOW DOWN
M11 - LANE BLOCK MERGE LEFT
M12 - SHARP TURN NEXT EXIT
M13 - TOLL ROUTE NEXT EXIT
M14 - ROAD WORK ALONG TURN US-81 WEST MOVE RIGHT
M15 - ROAD ALERT AUTO RADIO HWY-6 WEST STAY TUNED
M16 - ROCK SLIDE NEAR ROUTE HWY-6 EAST STILL OPEN
M17 - SLICK ROAD ALONG TURN I-835 WEST USE CHAINS
MESSAGES (Loaded)
140
TEST INSTRUCTIONS
Part A
(Subject car in Stationary Position of Predetermined Mark well out of
Legibility Distance)
Instructions to Subject - Legibility Study
The experiment that we want you to assist us with today consists of two
relatively simple parts, but will require some concentration on your part.
The first part involves the distance at which you can read the sign ahead of
you on the runway. In a moment I will ask you to proceed toward the sign at
a relatively slow speed, 10 to 20 mph. There will be a word displayed on the
sign. When you can read the word call it out to me. 1 will then ask you to
turn around and return to the starting point. We will make three of these
runs.
Are there any questions before we begin?
Part B
Instructions to Subje~ts - Unloaded Driver Study
The second part of our study today involves the reading of some typical
traffic messages. For this study you will be asked to drive toward and com-
pletely past the sign. From the starting point, we would like for you to
accelerate up to about 45 mph. As you proceed toward the sign a message
will come on. You will need to read and remember the message. Some of the
messages ~I/ill be displayed for a very short time, so be ready. Like~l/ise, some
of the messages will be followed by other messages, all of which you will need
to remember, so continue to observe the sign throughout the trip down the
run\·:ay.
After you pass the sign, stop at the two traffic cones. I will ask
you to repeat the message or messages to me at that time.
There will be a total of 16 runs. I will be in the car at all times
and will answer any questions that I can.
Before we begin, are there any questions?
Part C
Instructions to Subjects - Loaded Driver Study
This study is a little different from the one you ran the other dcy.
This time we would like for you to accelerate to 45 mph and drive through
the lane of cones you see ahead. While you are in the lane of cones the
sign will come on as before. The objective is to read the sign without
knocking down any cones. Each cone that you knock down represents one ac
cident that~ cause on the free;·:ay.
After you pass the sign, stop at the cones beyond and I 1t1il 1 ask you
to repeat the message. Some of the r;iessages will be displayed for a very
short time, so be ready. Likewise, some of the messages will be followed by
other messages, all of which you will need to remember, so continue to ob
serve the sign throughout the run through the lane of cones. Remember, read
the message to yourself and remember it, but do not knock down any cones.
Are there any questions before we begin?
142
Legibility Distance
Driver No.
ANS\~ER SHEET
0-F (Legibility and Unloaded)
Note: Message order depends on Driver No. --See attached list.
l. BOAT ______ _ 2. BOOK _____ _ 3. ROCK
Ml - ROCK SLIDE NEXT RIDGE
M2 SHARP TURN NEXT AHEAD
M3 - SHARP TURN NEXT RIDGE
M - SNOW SLIDE ALONG TURN 4
MS - SLIDE .11.REA ALONG TURN
M - TOLL STOPS NtXT i\.HEAD ; '6
M7 - TOU_ ROUTE NEXT EXIT H suer~ :) ('\ ·~ ~~ ALOf<G T: !D:1 I" ; ;J ··.11
8
Mg - AUTO v!RECf". t~.!_ o:--.; G ROAD ,, - TRUCK TOLL SLm.I omm I'll 0
Mll - ROC Y~ SL IDE r' r ., Cl i Cri J" :;ourE
Ml2 - LANE BLOCK JUST .DiH EAD
1., ' 13 - SLm·I TRUCK ALONG ROAD
Ml4 - TRUCK TOLL NEXT RIGHT
HWY-7 I :c~T i'i:_j I TOLL FREE
Pg. 2 Answer Sheet
-M --15 ROCK SLIDE ALONG TURN
US-23 EAST MERGE LEFT
Ml6 - SNOW SLIDE BEST ROUTE
H\~Y-6 ~JEST NEXT RIGHT
Ml7 - SLOW TRUCK ALONG ROAD
US-81 WEST USE BYPASS
144
ANSWER SHEET Driver No.
D-F (Loaded) --··---·-··-- --
Name
Legibility Distance
Ml - REST AREA ALONG ROAD 4------
Mz - TOLL ROUTE NEXT RIGHT
M3 - SNOW SLIDE AHEAD NEXT
M4 - ROAD vi ORK NEXT AHEAD
Ms - ROCK SLIDE ALONG TURN
Mr -0 SLOW TRUCK JUST AHEAD
M7 - LANE BLOCK NEXT EXIT
M,., - 1~UTO \·IRECK MERGE LEFT ''c) ------·--
Mg - TOLL STOPS SLOW Dmm
Mio- SLICK ROAD SLOW DOWN
M11- LANE BLOCK MERGE LEFT
M1r SH.l'\RP TURN NEXT EXIT
M13- TOLL ROUTE NEXT EXIT
Ml 4- ROAD ~iOR.K ALONG TURN
US-31 WEST MOVE RIGHT
145
Page 2 Answer Sheet
M15- ROAD ALERT AUTO RADIO
H\~Y-6 \·JEST STAY TUNED -
M15- ROCK SLIDE NEAR ROUTE
HWY-6 EAST STILL OPEN
M1r SLICK ROAD ALONG TURN
I-835 l~EST USE CHAINS
146
APPENDIX C
EXPERIMENTAL DESIGN
FOR
LABORATORY EVALUATION
OF
LM1P :'..ATRI/ 2 _::_ ' LOSS
EXPERIMENTAL DESIGN
TITLE
Effects of Bulb Loss on Legibility
OBJECTIVES
To determine for a given matrix, the percent of bulb loss that will
cause various alphanumeric messages to become illegible. This will
allow specifications to be determined for bulb replacement based on
percent of bulb loss.
FACILITY
Media Lab
TEST EQUIPMENT AND INSTRUMENTATION
0 Take slides of variable matrix sign
1 Cassette tn_pe of verb,°li instructions
SUBJECTS (Tota 1 Number -~] __ _)
Categories
Age
Sex
Education
TEST PERSONNEL A~lD SUPPORT
Experirnenter in lab to conduct test
TEST SCHEDULE
1- 5 Subjects per hour ~ession.
148
ORDER PRESENTATION i-10DE No. OF SLJB,JECTS --------
A 50%-10% 26
B so;s-1 mi, 25
,fJ., l 0%-50% 25
D 1 Q;{-50% 17 u
TEST DESIGN
Indepengent Variables
® Characters per word; size of matrix (random)
• Location of bulb failure (random)
® Percent bulb failure (10% increments)
Criterion Variables
® Percent correct response
Controlled Conditions
s Typ2 of presentation - sin~1e word flash
1 Presentation rate - 3 seconds per word
Statistical Design
I) 93 Subjects
o ; word lengths; 4 - 10 characters per word,
nu~ber length (4 characters - letters and numbers)
~ 40 words per study
• 5 levels of bulb failure per word (10% - 50%)
149
TEST DATA ANALYSIS
Data Reduction Methods
Calculate and plot percent correct response versus percent
bulb loss for various word lengths. Percent correct response
is defined as E
[l - NJ x 100
1t1here
E = Total of errors - either by omission or incorrect
reproduction.
N = Number of words presented.
ATTACHMENTS
e Detailed procedure and instructions
a Data sheets
150
FILM PR0CEDURE
Take slides of all 40 messages at 10% bulb loss in the following
seGuences:
& 5-4 character
e 5-5 character
fll 5-6 character
@ 5-7 character
(j 5-8 character
'ii 5-9 character
® 5-10 character
~ 5 - nurr:erals
Next, unscrew bulbs to simulate 20% bulb loss. Vary position C7
c!1?_rcctt~r \·;ord 1·:i i~~~in slices of all
Continue this seauence of a iticnal hulb loss to 50~. There w~~1
he ?~n total slidPS which will be 2rr~n=0~ in two rando~ orders usir~ lJO
s~~des in each order, and rrese~te~ in both descendinJ (SC~ - 1 (', ;~ '. IV-·;
WORD LIST
Four Character
s·1ow Toll Lane Road Exit
Five Character
Truck Alert Wreck Route Merge
Six Character
Bypass Access Bridge Reduce Median
Seven Character
Blocked Freeway Sta 11 ed Traffic Vehicle
Eight Character
Accident Entrance Dovmtown Pavement Junction
Nine Character
Condition
Ten Character
Congestion
Numerals
I-415
US-23
HWY-6
I-270
US-39
Diversion
Express\·1ay
Hazardous Alternate Co 11 is ion
Restricted Visib~lity Prohi!Jited
152
ORDER A
(sm; - 1 c . '
1 . Slow U)O) 24. Access (40)
')
l. . rccident (50) ')r C). Road ( 40)
--
3. Reduce (50) 26. Congestion (40)
4. Vi si bil ity (50) 27. Condition (40)
5. I-415 (50) 28. Pavement (40)
c:. Entrance (50) 29. JS-23 (40) v.
7. Route (50) 30. Entrance (40)
8. Road (50) '), J ! • Route (40)
9. Traffic (50) 32. .;;ccident (40)
10. Congestion (50) ".) ') .._;.). Sl O\•/ (40)
11. Alternate (50) 34. Reduce (40)
12. US-23 (50) 35. txoress 1t'lay ( 4" \ v;
13. Accc:.ss (50) "),; ...)\,). /\l ter-r18.te ( L:n l
' .. 1 J
14. Truck (50 37. Toll (~O)
15. E:xpr:-::s sway ( 50) ')r'l .)1j. Visibility (40)
16. Toll (50) '1Q .) ::J • 8"locked (40)
17. Blocked ( 5 ()) 4:J. Bypass ( 40)
18. Condition (50) 41. Congestion (30)
19. f:iyoass (50) 42. .L\ccess (30)
20. Pavement (50) 43. Road (30)
21. Truck ( Lrn J 44. Pavement (30)
"') c.. (_. Traffic (40) 45. US-23 (30)
I-:~l~~ 46. Entrance (30)
4 7 . Route ( 30)
48. Condition ( 30)
49. Bypass (30)
50. Express1;,1ay (30)
·-" ~1 (?Q) ::-J:. ):m•/ ,._;I_
52. Traffic (30)
53. I-415
54. Accident (30)
55. Alternate (30)
56. Toll (30)
57. Truck (30)
5 8 • v i s i b i 1 ity ( 3 0 )
59. Reduce (30)
60. Blocked (30)
61. Pavement (20)
62. Route (20)
63. Alternate (20)
611 .• Congestion (20)
6S. Road (20)
GS. Traffic (20)
67. US-2.3
68. Access (20)
69. Entrance (20)
70. Slm'I (20)
71. Bypass (20)
154
72. Accident (20)
73. Expressway (20)
74. Condition (20)
75. I-415 (20)
76. Truck (20)
77. Blocked (20)
78. Visibility (20)
79. Reduce (20)
80. Toll (20)
81. Route (10)
82. Expressway (10)
83. Entrance (10)
84. Slow (10)
85. Bypass (10)
86. I-415 (10)
87. Condition (10)
88. Traffic (10)
89. Road (10)
90. Access (10)
91 . Accident ( 1 0 )
92. Congestion (10)
93. Pavement (10)
94. US-23 (10)
95. Toll (10)
96. Blocked (10)
97. Alternate (10)
98. Truck (10)
99. Visibility (10)
100. Reduce (10)
155
ORDER B
(0::0) (50% - 10%) (40) l Vehicle 26. Merge I• ,'-1,J
2. !\ 1 ert (50) 27. Median (AO)
3. Restricted ( 50) 28. Lane (40)
4. Junction (50) 29. Collis ion (40) -
5. Hwy - 6 ( 50) 30. I-270 (40)
6. Exit (50) 31. Junction (40)
7. US-39 (50) 32. Alert ( £'1,Q)
8. Median (50) 33. Sta 11 ed ( £10)
0 ..lo Collision ( 5fJ) 3J.. Restricted (40)
10. Downtown (50) 35. Diversion r ,.,, \ ~ ..,_ \.) J
11. Freeway (50) 36. Bridge (tO)
12. Merge ( 50) 37. US-39 (40)
13. Lane ( 50) 38. \·:reek (40)
14. I-270 (50) ~q ..,~. Hazardous ( lQ)
15. Diversion (50) 40~ Vehicle (I' r': ' \ _,.,...) ;
16. Prohibited (50) 41. Prohibited (LC)
17. \!reek ( ~,--, ' JU i 42. I-270 c~s 1 v• I
18. Bridge rho) \ ._J. 43. DO\'in tm·m .! ; ("'; \
\ ....; .._, /
19. Hazardous (50) 44. Exit (30)
20. Sta 11 ed (50) c_::.. . ..J. Freeway '~n \ \ j.,j ,'
21. Exit (40) 46. Merge (30)
22. Freev1ay (40) 47. Diversion (30)
23. Prohibited ( llQ) 48. Hwy - 6 (30)
24. Hwy - 6 ( 40) 49. Bridge (30)
?r Do 1tinto1,;ri I 1ln l 50. Junction (30) -J- \ .· ·'
156
51 . Lane (30) 76. US-39 (20)
52. Sta 11 ed (30) 77 .. Free't-1ay (20)
53. Res tr"icted (30) 70 CJ. Lane (20)
sc... ~-ire ck (30) 79. Sta 11 ed (20)
55. US-39 (30) 80. Hazardous (20) -
56. Hazardous (30) 81. Median ( l 0)
57. Median (30) 82. US-39 ( 10)
58. l/ehic'!e (30) 83. Junction ( 10)
59. Collision (30) 82- Merge ( 1 0) . ,,
60. Alert (30) p::; Vehicle I., ,...,, \ vv. \ i !J j
61. Dm·mtovm (20) n- Diversion ( i C1 \ Qh ._,. \ I I j
62. Exit (20) 87. I-270 ( 10)
63. Vehicle (20) 82. Exit ( 10)
Git. I-270 (20) 30 Bridge I.,. r \ J. \ l u J
65. Diversion (20) 90. Prohi biteci I~ ,..... '\
(~ ! 0)
66. \fr eek (20) 91. Dm·mto'.·m ( 1 0)
67. Junction (20) 92. \·!reek ( 10)
58. , ... (?nl 0 ') flc:zardoc.is ! , ~ '1 :;ed1an ·.·-·_,I _; .) . \ l u j
f.9. ,Junction ( 2 ()) nft Hv1~l - 6 ( 1 'J) ::;~
I(\ ~v1y - c. (20) 95. Lane ( 1 0) Iv. c.
71. /\lert Ir" ) \ ( i j
f' --:::-0,. Alert ( 1 ,-, ' •j j
7? I ._ .. Bridge (20) 97. Sta 11 ed ( l 0)
73. Collision (20) 98. Restricted (10)
7 !l. Merge (20) 99. Collision ( 10)
75. Restricted (20) 100. Free~-1ay ( l 0)
157
ORDER A
1. Reduce ( 10) ( ·1or -- ,.- ")Cf) J1..J ,.., 26. I-415 (20)
2. Visibility ( 10) 27. Condition (20)
3. Truck ( 10) 28. Expressviay (20)
4. Alternate (10) 29. Accident (20) -
5. Blocked ( 10) 30. Bypass (20) .
6. Toll ( 10) 31. Slow (20)
7. US-23 ( 10) 32. Entrance (20)
8. Pavement (10) 33. Access (20)
9. Congestion ( 10) 34. US-23 (20)
10. Accident ( 10) 35. Traffic (20)
11. Access ( 10) 36. Road (20)
12. Road ( l 0) 37. Congestion (20)
13. Traffic ( l 0) 38. . .l\ lternate (20)
1 t,. Condition ( 1 0) 39. Route (20)
15. I-415 ( 1 0) 40. Pavement (20)
16. Bypass ( l 0) 41. Blocked (30)
17. S 1 O':/ I., '"' \ \ I '-' / 42. Reduce (30)
18. Entrance ( 10) 43. Vis i bi 1 ity (30)
19. Express1·1ay ( 1 0) L'A. Truck (30)
20. Route ( 10) 45. Toil ( 30)
?1 ~I. Toll (20) 46. Alternate (30)
22. Reduce (20) 47. Accident (30)
23. Vis i bi 1 ity (20) Ll,8. I-415 (30)
24. Blocked (20) 49. Traffic (30)
?::; ---'· Truck (20) 50. s l Qi•/ ( 30)
158
51. Expressway (30) 76. Road (40)
52. Bypass ( 3()) 77. /1.ccess ( i"cQ )
J;:{ ..;..,,. Condition ( 3()) 78. I-415 (40)
54. Route (30) \ l_,, 79. Traffic (40) -
55. Entrance (30) 80. Truck (40)
56. US-23 (30) 81. Pavement (50)
57. Pavement (30) 82. Bypass (50)
58. Road (30) "~ Condition 1-,-... ~: "'i \ O'J) v-.
59. Access (30) 8d. Blocked r i::o ~ \ . ...,.; I
60. Congestion (30) 85. Toll (50)
61. Bypass ( 4-0) 86. Express1>1ay (50)
62. Blocked (40) 87. Truck (50)
63. Visibility (40) 88. Access ( 50)
64. Toll (40) 89. US-23 (50)
65. Alternate (40) 90. P., 1 tern ate ( 50)
66. Expressway ( ~-0) 91. Cornestion (50)
cl. R.e(iuce (en) 92. T:-2Fi c ( 5Q;
;:o u'-: .. s 101'/ (4Q) C' ')
':).). Road (50)
69. l\cc·i dent (40) 94. Route (50)
7n r V •
Pn1d-a i\•..1\...i. L..\... (40) j'J. ~r:trance
,' -r·. '· ,,:::!'._j:
71. Entrance ( 40) 96. J-l'i.15 (50)
72. US-23 (40) 97. Visibility (50)
73. Pavement (40) 98. Reduce (50)
74. Condition ( ,~ 0) ,..,. 99. Jkc i dent (50)
75. Congestion ( 40) 100. Sl O\•I ( 50)
159
ORDER B l. Freeway ( l 0) ( i o~~ - so~o 26. Restricted (20)
2. Collision ( 10) 27. Merge (20)
3. Restricted ( 10) 28. Collision (20)
4. Sta 11 ed (10) 29. Bridge (20) -
5. Alert ( 10) 30. Alert (20)
6. Lane ( 10) 31. Hwy - 6 (20)
7. Hwy - 6 ( 10) 32. Junction (20)
8. Hazardous ( l 0) 33. Median (20)
9. \•ire ck ( 10) 34. Prohibited (?n\ \. ,_ . ._, J
10. Dovmtown ( 1 0) .... ,.. Wreck (20) j::J.
11. Prohibited ( 10) 36. Diversion (20)
12. Bridge ( 1 0) '? -....,; . i-270 (20)
13. Exit (10) 38. Vehicle (20)
14. I-270 ( 10) 30 J. Exit (20)
15. Diversion ( 10) ll.Q. Dat'lntmm (20)
16. Vehicle (10) 41. Alert (30)
17. l!erge ( 1 n) 42. Collision (30)
18. Junction ( 10) {! ?. Vehicle ( 30 l ,~...;. j
19. US-39 ( 10) 44. Median (30)
20. r~~d i an ( 10) i_::._ Hazardous (30)
21. Hazardous (20) 46. US-39 (30)
22. Stalled (20) 47. Wreck (30)
23. Lane (20) 48. Restricted (30)
24. Freev1ay (20) Ll,9. Stalled (30)
25. US-39 (20) 50. Lane (30)
160
5 i . Llunction (30) 76. Downtovm (40)
·-r ~) ;~ . ~2ridge (30) 77. µ,.n,
•• "j' - 6 (40)
C:,".; Hv1_y - r::. (30) 78. Prohibited (40) \~
,.- .~ Diversion ( .30) 79. Freeway (40) .,_i ... :-.
-
55. Mer9e ( 30) 80. Exit (40)
56. Freeviay (30) 81. Stalled (50)
57. Exit (30) 82. Hazardous ( 50)
58. Dovmtown ( 3 ()) ~ .... Bridge (50) b-5.
59. I-270 (30) 81 ';!reek ( 50)
60. Prohibited (40) ,..,~ Prohibited f _,... ', :::'.::l. \2'.J)
61. Vehicle (40) 86. Diversion (50)
62. Hazardous ( 40) 87. I-270 ( ~n \ \ ..... ...,; J
63. \1/reck (40) 88. Lane (50)
6~-. US-39 ( 110) 89. ~1erge (~'I\ ..., -... /
f. 5. Bridc;2 ( L:-0) Q"' _.u. Freev1ay ! : (\ \ -. :)'J /
6G. f!iversion ( 40) 91. ~ovmtc'.-.:n I r-,... \ p ...... : \ l
\~·~I
(:,: -, F:~~s tri cT~ed ft--~ -.i 92. Collisicn ( ::J ': . ,_,,
EC. Sta 11 ec'. ( ilQ) 93. Median ( c:: () ' " ...,. -· I
6?.. J\lert ( '-"'O) C·'i _,~. US-39 (-" \ :iu;
7n .Junctirin ( llQ) r- r :j :J • Exit (50)
7L I-270 ( 4'J) 96. ~h·;y - 6 (50)
72. Coll is ion ( 40) 97. Junction ( rn \ 0U j
73. Lane (48) 98. Restricted (50)
74. Median ( 11-'J) 99. Alert (50)
7 :. ~ Merrie ( 40) 100. Vehicle (50)
161
r·~anua 1 Sta rt
Auto Slide Advance and Projector On
Auto Slide Projector Off
Auto Tape Stop
TEST PROCEDURE AND INSTRUCTIONS
This is a study of bulb loss on an electrical message
display sign similar to this:
(Slide of Matrix Sign with Message)
The objective of the study is to determine how many
bulbs could be out on this sign and the message still
be readable to the driver.
We will flash various single word messages of different
lengths on the screen before you .. The message may or may
not be readable, as there may be only a few bulbs on or
all the bulbs on the word. Each word will be shown fer
a few seconds and then flashed off. If you can read the
word, write it next to the corresponding blank on the
answer sheet. If you cannot read the \'lord, mark an "X"
in the corresponding blank on the answer sheet.
To help you keep track and be ready, I will announce each
word message as it appears on the screen. Now, are there any questions?
All right, the first word will appear in just a second.
Remember, when each slide goes off, if you can read the word,
write it on the answer sheet. If you cannot read the word, fi'lark an "X" on the answer sheet.
(Pause) 162
Auto Slide P.dvance and Projector On
Auto S"lide Projector Off
Auto Slide Advance and Projector On
Here is the first message:
( Fi rs t Message)
(3 seconds)
( l 0 second pause)
Here is message number 2:
Here is the last message:
coc'pletes this
Initial Instructions
Q u e s t I •Jr1 s .
Messages .
Lost Ti:;:e.
Total Time.
40 \·1ord messages
2 random word lists
part of the study. Thank ~/JJ very
25.0 minutes
0 .. 5 rninutes
30. 0 minutes
2 nodes of presentation - ascending, descending
E3
Word Message
1.
2.
3.
4.
5.
lOG
ANSvJER SHEET
Respondent # --
164
>--' CTI U-1
ORDrn
Characters _ _rs;r Word No. of \fords
--------- --·· ... __ (_NJ_ ________ - -
4 C/W
5 C/W
6 C/W
7 C/W
8 C/W
9 C/vJ
l 0 C/vJ
Numerals
PRESENTATION MOOE
Respondent # ______ _
~~ Bulb Loss No. of Errors % Cor:-rect Re_~ons_e __
Omission ( E) Wrong j ( 1-E/l!_)_Ll_O_Q_ __ _
APPENDIX D
EXPERIMENTAL DESIGN
OF
LABORATORY EVALUATION
OF
SYMBOLIC MATRIX SUBSTITUTION
166
EXPERIMENTAL DESIGN
TITLE
Sy1nbo 1 ogy Study
OBJECTIVES
To determine the correlation between symbols formed by a matrix
sign and painted signs.
FACILITY
Media Lab
TEST EQUIPMENT AND INSTRUMENTATION
8 Slides of matrix and static symbols
® Instructions
@Slide projector
8 AnsvJer sheets
0 Tables and chairs
SUBJECTS (Total Number 50
Categories:
.L\ge
Sex
Education
TEST PERSONNEL AND SUPPORT
One Administrator required
TEST SCHEDULE
1-10 subjects per test until 50 subjects obtained
167
TEST DESIGN
Independent Variables
Symbol presentation
Criterion Variables
Percent correct response
Controlled Conditions
Presentation time
• Test symbol 3 seconds
© Choice symbols -- 5 seconds
Statistical Design
50 subjects
13 symbols
TEST DATA ANALYSIS
Data Reduction Methods
Calculate ~ correct response ~ersus symbol type
ATTACHMENTS
• Test Symbols
i Detailed Procedure and Instructions
• Data Sheets
168
STUDY D-3
SLIDE ORDER
l. EA (Figure 24)
2. EB (Figure 24)
3. lA ( Symbo 1 III-1 -- Figure 21 )
4. 18 (Symbol I I I-1 -- Figure 21)
5. 2A (Symbol I-2 -- Figure 19)
6. 28 (Symbol I-2 -- Figure 19)
7. 3A (Symbo 1 III-2 -- Figure ? 1 \ ;.._ l /
8. 38 (Symbol III-2 -- Figure 21 I I )
9. 4A ( Symbo 1 IV-2 -- Figure 22)
10. 48 (Symbol IV-2 Figure 2?' -J
11. 5A (Symbol II-1 Figure 20)
12. 58 (Symbol II-1 Figure 20)
13. 6A (Symbo1 I-4 -- Figure l 9)
14. 68 (Symbol I-4 -- Figure lQ\ 'JI
15. 7A (Symbol II-2 Fig~:e ......,,....,, -- {_ 0)
16. 78 (Symbol II-2 -- Figure 20)
17. 8A (Symbol I-3 Figure 19)
18. SB (Symbol I-3 Figure 19)
19. 9 ,D, (Symbol I-1 Fiugre 1 9)
20. 9B ( Symbo 1 I- l Figure 19)
21. l CJA (Symbo 1 IV-1 Figure ?? \ --}
22. lCS ( Symbo ·1 IV-1 Fi g'..lre 2~' , ' t- j
23. llA (Symbol IV-3 Figure 22)
24. 11 B ( Syr:ih." 1 •• !....... ... ) IV-3 Figure ??" --)
169
MATRIX AND GRAPHIC SYMBOLS
170
SYMBOL 1
A B c D
171
SYMBOL 2
A B c D
172
SYMBOL 3
A 8 c D
173
SYMBOL 4
A 8 c D
174
SYMBOL 5
D A B c D
175
SYMBOL 6
A 8 c D
176
SYr<1BOL 7
A B c D
177
SYMBOL 8
A B c D
178
SYMBOL 9
A B c D
I/,_!
SYMBOL 10
A B c D
180
SYMBOL 11
-1;
D i'
A B c D
181
SYMBOL 12
A 8 c D
182
SYMBOL 13
A B c D
12.3
Auto Tape Stop
Manual Start
Auto Projector On
Auto Slide Advance
Auto Projector Off
TEST PROCEDURE AND INSTRUCTIONS
This is a study to determine the similarity
betv1een symbols painted on a highway sign, and the
same symbol formed on an electronic display sign,
similar to the one you see on the ~creen before you.
We will flash various pairs of slides on the screen.
The ftrst slid~ in eath pair will show a symbol that
might be flashed on an electronic display sign. The
second slide in each pair will show four painted
symbols similar to the symbol shown on the first
slide. On the answer sheet before you, we would
like you to circle the letter of the corresponding
symbol most resembling the symbol shown first.
The length of time that each slide will be on will
be relatively short, so watch carefully. To help
you get accustomed to the procedure, we will be
rresent an example pair of slides first. Here is
the example pair of slides:
(3 seconds)
(Slide EA)
(5 seconds)
(Slide EB)
To help you keep track and be ready, I will announce
number of each pair of slides as it appears on
the screen. Now, are there any questions?
184
Auto Tape Stop
Manual Restart
Auto Projector On
Auto Slide Advance
Auto Projector Off
Auto Slide Advance
Auto PrGjector On
All right, the first pair of slides will appear
in just a second. Remember, circle the letter "A,"
"B," "C," or 11 0" corresponding to the symbol on the
second slide most like the symbol sho1tm on the first slide.
(Pause)
Here is the first pair of slides:
(3 seconds)
(Slide lA)
(5 seconds) (Slide lB)
(15 seconds)
Here is slide set =2.
(3 seconds)
(Slide 2A)
And this is the last pair of slides.
135
Auto Projector On
Auto Slide Advance
Auto Projector Off
Auto Tape Stop
(3 seconds)
(Slide l 3A)
(5 seconds)
(Slide l 3B)
(15 seconds)
OK, that completes this study. Thank you very much.
Initial Instructions 3.0 riinut2s Questions 2.0 r;;i nutes Symbols. 5.5 minutes Lost Time 0.5
. ._ m1nut..es Tota 1 Time 11. 0 :ninutes
186
ANSvlER SHEET
Narne Time Date ··-----·---------------------- ----
(Ci re 1 e Your Answer)
Example A B c D
1. A B c I\ cJ
2. A B c D
3. A B c D
4. A 8 c 'l L'
5. A B c D
6. f'\ B c D
7. p 'l u c D
0 u. A B c D
9. A B c D
l 0. A B c D
11. A B c D
12. A B c D
187