Professional Documents
Culture Documents
Twenty-five internationally proposed symbols for fire-safety alerting were evaluated for un-
derstandability by 91 U.S. participants. Three modes of symbol presentation (slides,
placards, and booklets) and two modes of participant response (definition and multiple
choice) were studied. Confidence ratings and production data (drawings) were also obtained.
Mode of symbol presentation had no significant effect, while definition and multiple-choice
response procedures led to generally similar conclusions. Confidence ratings were useful in
reconciling discrepancies between the two response methods. The understandability of the 25
symbols ranged from near zero to virtually complete comprehension. The poor performance
of some critical symbols such as "exit" was noted, and some potentially dangerous confu-
sions in meaning were revealed.
eating "exit" and "not an exit," ISO has since Walker et al. (1965) used placard-type sym-
responded by replacing these with new boIs. Griffith and Actkinson (1978), Dewar
symbols. and Ells (1977), and Cahill (1975) used col-
The present paper reports an experiment ored slides to present their stimuli. Still other
on the understandability of the new ISO researchers (Wiegand and Glumm, 1979;
egress symbols, together with other exit desig- Green and Pew, 1978; Easterby and Zwaga,
nations and the remaining set of proposed 1976; and Freedman, 1978) used booklets or
ISO fire-safety symbols. The purpose of the reproductions that allowed subjects to work
experiment was twofold: (1) to determine the at their own pace. In no case, however, did
understandability of each proposed symbol to any researcher compare a symbol's perfor-
a group of potential users, and (2) to assess mance across the various stimulus presenta-
the influence of certain methodological vari- tion modes.
ables in measuring symbol comprehension.
METHOD
Because a symbol must communicate its
intended meaning to the target audience, un- In the present study, the relative effective-
derstandability is one of the most important ness of three presentation modes and two re-
criteria for determining a symbol's effective- sponse methods was compared in an evalua-
ness. Understandability or meaningfulness tion of the understandability of 25 fire-safety
has been assessed in many different ways. symbols. The experiment consisted of a main
Perhaps the most frequently used method has portion on symbol meaningfulness and a
been to ask participants to provide short supplementary one on symbol production.
definitions for each member of a set of sym- The meaningfulness portion evaluated how
bols (Brainard, Campbell, and Elkin, 1961; well a set of fire-safety symbols conveyed the
Walker et aI., 1965; Cahill, 1975; Easterby intended messages. The production portion
and Zwaga, 1976). Wrong answers given in a assessed the kinds of images that participants
definition procedure can provide insight into produced for a given message.
the kinds of confusions and misunderstand- The procedure of the symbol meaningful-
ings that are associated with a particular ness portion (Part 1) made use of a two-factor,
symbol. Still other researchers have asked independent-groups experimental design.
participants to select a definition for each One factor was the mode of symbol presenta-
sym bol from a long list of definitions tion (placards, slide projections, or booklets);
(Brainard, et aI., 1961; Griffith and Actkinson, the other factor was the type of response re-
1978; Wiegand and Glumm, 1979). Various quired of the subject (providing a short defi-
other rating, ranking, and behavioral mea- nition of the meaning of each symbol or
sures have also been used to evaluate symbol selecting the correct meaning from four
meaningfulness (Forbes, Gervais, and Allen, multiple-choice alternatives).
1963; Dewar and Swanson, 1972; Easterby
Participants
and Hakiel, 1977; Easterby and Zwaga, 1976;
Dewar and Ells, 1977; Green and Pew, 1978), A total of 91 paid participants took part in
although comparisons of performance across the experiment. They were screened to
the various response methods have rarely exclude those reporting any visual problems
been made. (other than corrective lenses). The partici-
In the investigations of symbol effective- pants included 58 women (range 18 to 72
ness, researchers have used a variety of years; median age, 41) and 33 men (range 18
stimulus materials. Brainard et al. (1961) and to 63 years; median age, 25).
feet of the mode of symbol presentation or of ing of the definition data yielded generally
the interaction of this factor with the type of similar information: the rank-orders of the 25
response. The two response methods differed symbols under the two scoring criteria were
significantly only when the definition data in good agreement (p = 0.96, p < 0.001).
were strictly scored. Strict and lenient scor- Table 1 (left side) lists the percentage of
EXIT
Emergency Exit, Not an Exit Not an Exit, U.S. Exit
Currently Currently Originally Proposed Two Versions Tested:
Proposed Proposed (Black on Yellow) Green on White
(White on Green) (White on Black, Red on White
Red Circle & Slash)
Figure l---Continued
correct answers for each symbol, for both the ods. with the correlation coefficient being
multiple-choice and definition groups. (Data 0.86 (p < 0.001) for lenient scoring and 0.87 (p
are combined across method of presentation.) < 0.001) for strict scoring.
Correlational analysis reveals general Table 1 (right side) contains the mean
agreement between the two response meth- multiple-choice confidence ratings for each
TABLE 1
Percentage of Correct Answers and Mean Confidence Ratings for Each Symbol
Confidence Ratings
Multipfe choice
Percent Correct
Symbol Highest
Multiple Definition Definition Correct Correct Rated
# Meaning Choice (Lenient) (Strict) Answers Answers Alternative
All Correct
Subjects Subjects
1 Fire Extinguisher 100.0 97.8 97.8 4.7 4.7 1.4
2 Hose & Reel 95.7 91.1 77.8 4.8 4.8 1.9
3 Fire ladder 19.6 53.3 24.4 2.9 4.1 3.8
4 Fire Bucket 95.7 80.0 60.0 4.6 4.6 1.8
5 Fire Fighter's
Equipment 0 2.2 0 1.5 4.0
6 Direction to
Equipment 22.2 6.7 2.2 2.5 4.0 3.2
7 Break Glass
For Access 28.3 15.6 11.1 2.8 4.2 3.0
8 Slide Door
To Right 56.5 15.6 4.4 3.3 4.1 2.8
9 Do Not Use
Water 76.1 91.1 88.9 4.3 4.8 2.5
10 Do Not Lock 89.1 84.4 68.9 4.6 4.8 1.7
11 No Smoking 100,0 95,6 95.6 5.0 5.0 1.0
12 No Open Flame 87.0 88.9 75.6 4.1 4.5 1.8
13 Do Not Block 2.2 20.0 0 1.6 5.0 4.0
14 Keep Fire Door
Shut 18.5 35.6 28.9 2.3 3.7 3.4
15 Emergency Phone 93.5 93.3 71.1 4.3 4.4 1,5
16 General Phone 87.0 95.6 82.2 4.6 4.8 1.6
17 Fire Alarm Horn 77.8 22.2 22.2 3.2 3.5 1.8
18 Fire Alarm
Call Point 32.6 15.6 13.3 2.6 3.7 3.1
19 Fire Exit-
(Yannone. 1979) 95.7 91.1 86.7 4.7 4.8 1.8
20 Emergency Exit 57.8 62.2 22.0 2.5 3.2 2.2
ISO Original
21 Emergency Exit
ISO Proposed 69.6 86.7 68.9 3.8 4.5 2.6
22 No Exit
150 Proposed 69.6 53.3 33.3 4.0 4.5 2.1
23 No Exit
ISO Original 30.4 11.1 6.7 2.0 3.5 2.8
24 U.S. EXIT-Green 93.5 97.8 93.3 4.7 4.8 1.3
25 U.S. EXIT-Red 91.3 97.8 97.8 4.8 5.0 1.6
Mean = 3.6 4.3 2.4
TABLE 2
ISO Images and Most Frequently Produced Images for Each Referent
Group mean confidence ratings tended to than did subjects under 50. There was no sig-
be higher for those symbols that were most nificant effect of sex.
often correctly identified. The correlation Table 2 summarizes the results of the pro-
coefficient of group mean ratings for the cor- duction procedure, listing the most fre-
rect multiple-choice alternative with the per- quently produced image content for each of
centage correct is r = 0.95. The correct choice the 12 referents. For comparison, the pro-
alternative usually received the highest group posed ISO image content for each referent is
confidence rating; however, in eight of the 25 included in the table. Participants were spe-
cases, one of the incorrect alternatives re- cifically instructed to avoid using words; de-
ceived a higher confidence rating. These same spite this, words were among the most com-
alternatives frequently appeared as incorrect mon responses for several referents. These
answers from the definition groups. referents (and responses) included Exit and
The correlation coefficient of age with Fire Exit ("Exit"), Hazard Area ("Danger"),
number of items correct (lenient scoring, and Blind Alley ("No Exit," "Dead End,"
collapsed across stimulus presentation "Stop").
method and type of response) was r = -0.30,
DISCUSSION
which, while significant (p < 0.01), accounted
for only about 9% (r2 = 0.09) of the variance in The results of the present experiment indi-
the scores. Subjects over 50 years old aver- cate that mode of symbol presentation
aged about two fewer items correct (out of 25) (placards, slides, or booklets) had no dis-
low confidence ratings, suggesting that tegral component of the design. Since the re-
multiple-choice participants guessed at the sponse to the arrow alone and the results
correct alternative. Thus, the ratings provide from the production data both indicated that
a check on guessing for multiple-choice data. an arrow may imply egress, the superior
Regardless of the response measure used, performance of this third exit symbol may be
the 25 symbols differed widely in understand- attributable to the integral arrow.
ability, with percentage-correct measures Problems also arose for the two "not an
ranging from zero to 100% and mean confi- exit" symbols. The newer symbol, showing a
dence ratings for correct alternatives ranging figure, open door, circle, and slash, was fre-
from 1.5 to 5.0. It should be noted, however, quently interpreted as "do not run" or "do
that while the total of 91 participants can not enter." The originally proposed symbol.
provide a good initial assessment, definitive showing a triangle, square, and arrow, re-
statements require much more extensive and ceived very few correct answers and particu-
representative sampling. larly low confidence ratings. Much more im-
Because some of the fire-fighting equip- portantly, potentially dangerous misinterpre-
ment symbols intended to be used with a di- tations were common for this symbol: "exit,"
rectional arrow appear to be relatively "shelter area," "elevator," and "go this way"
meaningless, the arrow was tested alone to accounted for 40% of all answers, while
determine if it had a meaning itself. In fact, "overhead hazard" accounted for another
the arrow conveyed the idea of "exit," "one 23%. Such confusions represent the most
way," or "go this way" to 82% of the partici- dangerous failure of a symbol as an element
pants. These data suggest that any use of an in a building communication system. Similar
arrow as a directional indicator for equip- dangerous confusions emerged for "fire lad-
ment should be clearly integrated with the der" (interpreted as fire escape or egress) and
equipment symbol to minimize false conno- "fire fighter's equipment" (interpreted as
tations of egress. shelter or tunnel). In each case, misinterpre-
When specific symbol performance is con- tation of the symbol could send an occupant
sidered in detail, the most important fire- seeking egress in the wrong direction.
safety symbols to be discussed are those relating Data from the supplementary production
to safe egress. Three exit-related symbols, procedure indicate that in only 5 of the 12
two word signs, and two "no exit" symbols cases was the image content essentially the
were included in this experiment. Two same as that proposed by the International
exit symbols proposed by ISO for emergency Organization for Standardization (1978), and
exit did not test well. One, a white rectangle in two of these five cases, another alternative
on a green square, appeared to have little in- was a close second. Although participants
trinsic meaning, judging by the small per- had viewed images for these referents in Part
centages of correct responses and generally 1, this did not appear to substantially bias
low confidence ratings. The other, a more re- their response toward producing the ISO
cent ISO symbol, showing a stylized human image. Where the production image content
figure stepping through an open door, per- differed from the ISO image content, it
formed better but still suffered by implying tended to be for symbols that had performed
the need to run. Although another exit symbol poorly in the understandability test, such as
showing a figure, an arrow, and a fire tested "fire alarm call point," "exit," "not an exit,"
well, it included a directional arrow as an in- and" do not block."