You are on page 1of 10

CHAPTER

4.

TEXT. GRAPHICS.

SYMBOLS.

AND CODES

119

format of part 2 would be better. The that the value for condition b is more than three times a because we "see" height. Part 4 corrects the distortion (1983) provides many such examples useful guidelines that would enhance played.
SYMBOLS

common perception of part 3 probably is than twice that for . and that c is more the volume of the block rather than their by showing the three heights only. Tufte of perceptual distortions and provides the correct perception of the data dis-

Civilization abounds with a wide assortment of visual symbols and signs that are intended to convey meaning to us. Examples include symbols for men's and women's rest rooms (which often are less distinguishable from one another than we would like them to be); symbols for windshield washers and wipers Q.fI some automobile dashboard controls; and little icons presented on computer screens of trash cans, file folders. and filing cabinets tosymboliz.e various computer operations.
Comparison of Symbolic and Verbal Signs

Where there is some question as to whether to use a symbolic or verbal sign, a symbolic sign would probably be preferable if the symbol reliably depicted visually what it is intended to represent. One argument for this (supported by some research) is that symbols do not require the that words or short statements do. For example, a road sign showing a deer conveys immediate meaning, whereas use of the words requires the recoding from words to concept. Note. however. that some symbols do not visually resemble the intended concepts very well and thus may require \earning and recoding. A study that supports the possible advantage of symbols conducted by Ellis and Dewar (1979) where subjects listened to a spoken traffic message (such as "two-way traffic") and were then shown a slide of a traffic sign (either a symbolic sign or a verbal sign) and asked to respond yes or no, depending on whether they perceived the sign as being the same as the spoken message. This was done under both visually "riondegraded" and "degraded" conditions. The mean reaction times, in Figure 4-15, show a systemic superiority for the symbolic signs, especially under the visually degraded conditions.
Objectives of Symbolic Coding Systems

When a system of symbols is developed, the objective is to use those symbols that best represent their referents (that is, the concepts or things the symbols are intended to represent). This basically depends on the strength of tion of a with its This association, in turn, depends on either of two factors; any already established association [Cairney and Siess (1982) call this and the ease of learning such an association.

120

PART 2: INFORMATION INPUT

1000

". 900 i!
o
v

c Q
v

~ 800-

700'-:----'---,.--,.----Nonoeqraded

---::--.... 1--,-Degraded

FIGURE 4-15 Mean reaction times of yes and no responses to symbolic and verbal traffic signs viewed under nondegraded and degraded viewing conditions. (Source: Ellis and Deward. 1979. Fig. 1. p. 168.)

At a more specific level. in Chapter 3 we discussed certain guidelines for using coding systems: detectability, discriminability, compatibility, meaningfulness, and standardization. These guidelines are equally applicable to the design of symbol systems. After all, symbols are really just another form of coding.
Criteria for Selecting Coding Symbols

Some coding symbols already exist (or could be developed) that could be used with reasonable confidence. If there is any question about their suitability. however, they should be tested by the use of some experimental procedure. Various criteria have been used in studies with such symbols. A few criteria are discussed below: Recognition On this criterion. subjects usually are presented with experimental symbols and asked to write down or say what each represents. Matching In some investigations. several symbols are presented to subjects along with a list of all the referents represented, and subjects are asked to match each symbol with its referent. A variation consists of presenting the subjects with a complete array of symbols of many types and asking them to identify the specific symbol that represents each referent. Sometimes reaction time is measured as well as the number of correct and incorrect matches.

CHAPTER

TEXT, GRAPHICS,

SYMBOLS,

AND CODES

121

Preferences and Opinions In other circumstances people are asked to express their preferences or opinions about experimental designs of symbols.
Examples of Code Symbol Studies

To illustrate examples.

the procedures

used In code symbol studies,

we give a few

Mandatory-Action Symbols The first example deals with a set of so-called mandatory-action messages and their respective symbols, illustrated in Figure 4-16 (concerning use of protection devices for the ear, eye, foot, etc.) (Cairney and Siess, 1982). These symbols were shown to a group of newly arrived Vietnamese in Australia, who were first told where the signs were to be used and then asked to state what they thought each sign meant. A week later the subjects were given a recall test in the same manner. Figure 4-16 shows, for each symbol, the percentage of correct responses for both the origina! test (O~ and for the second test (R). The results of this particular study show rather dramatically how well a group of people with a different cultural background could learn the intended meaning of those particular symbols. (Only the first symbol did not come through very well.) This example illustrates the use of a recognition criterion and the desirable objective of using symbols that are easily learned. even if the symbols are not initially very recognizable. Comparison of Exit Symbols for Visibility When a -specific symbol has a poor association with its referent, there can be a problem in selecting or designing a better symbol. One approach is to create alternative designs for the same referent and to test them. Such a procedure was followed by Collins and Lerner (1983) with 18 alternate designs of exit signs, particularly under difficult viewing conditions simuFIGURE 4-16

Symbolsof mandatory-action messages used in a study of recognition and recallof such symbols.The percentagesbelowthe symbols are the percentages ofcorrect recognition, as follows: o = originaltest; R = recan1 week later. (Source: Adapted from Cairney and Siess, 1982,
Fig. 1.)

.O()f)i)O
1: Must use ear protection 2: Must use eye protection 3: Must use foot protection 30% 100 4: Must use hand protection
50%

5: Must use head protection 37% 97

6: Must use breathing protection 13"" 97

0
R

10% 73

20% 96

97

122

PART 2:

INFORMATION

INPUT

lating an emergency situation. (After all, you would like to be able 10 identify the exit if a building were on fire!) The subjects were presented with 18 designs of exit symbols 'under different levels of viewing difficulty (with very brief exposure time) and asked to indicate whether each was. or was not. an exit sign. Certain of the designs are shown in Figure 4-17. along with their percentages of error. (In reviewing the errors. it was found that certain symbols for "No exit" were confused with those for "Exit.") Certain generalizations about the features of the best signs can be made: (1) "filled" figures were clearly superior to' "outline" figures, (2) circular figures were less reliably identified than those with square or rectangular backgrounds. and (3) simplified figures (as by reducing the number of symbol elements) seem beneficial.
Perceptual Principles of Symbolic Design

Much of the research regarding the design of symbols for various uses has to be empirical. involving. for example, experimentation with proposed designs. However. experience and research have led to the crystallization of certain principles that can serve as guidelines in the design of symbols. Easterby (\967. 1970), for example. postulates certain principles that are rooted in perceptual research and generally would enhance the use of such displays. Certain of these principles are summarized briefly and illustrated. The.illustrations are in Figure 4-18. Although these particular examples are specifically applicable to machine displays (Easterby, 1970), the basic principles would be applicable in other contexts. Figure to Ground Clear and stable figure-to-ground articulation is essential. as illustrated in Figure 4-18a. In the poor, unstable figure, the direction in which the arrow is pointing is ambiguous.

Green % error -10

& White

Black & White

Green & White

AGURE 4-17 Examples of a few of the 18 exit signs used in a simulated emergency experiment. with percentages of errors in identifying them as exit signs. (Source: Adapted from Collins and Lerner. 1983.)

. f'

,
. .

Red. White & Black


0/0

Black & White

Black & White

error -+-

39

40

42

---------------------------~. --.-

--.. -.~.

----

CHAPTER

TEXT. GRAPHICS.

SYMBOLS. AND CODES

123

"'(ED
Good. stable figure (b) Figure boundary

(d) Slmpltcrty

Poor. unstable

figure

(c) Closure

I
mi~ . Closed figure readily perceived

~Ie
A srmple shape IS readtlyl perceived

(e) UnIty

~
Good

c::::>
Poor Poor
__ L .
-L-

@
1

I
I
Too much detail makes a weak symbol

All parts of trns symbol he wllhin a single boundary making a good symbol

.,

=1 \ x, m/mrn

Bxmimin

Deiau outside

makes

a _

I__ o_p_en_fi_gu_r....:e_-'has less impact

I rx;o;. symbol ~

FIGURE 4-18 Examples of certain perceptual principles relevant to the design of visual code symbols. These particular examples relate to symbols used with machines. (Source: Adapted from Easterby,
1970.)

Figure Boundaries A contrast boundary (essentially a solid shape) is preferable to a line boundary, as shown in Figure 4-18b. Closure A closed figure, as illustrated in Figure 4-18c, enhances the perceptual process and should be used unless there is reason for the outline to be discontinuous. Simplicity The symbols should be as simple as possible, consistent with the inclusion of features that are necessary, as illustrated in Figure 4-18d. Unity Symbols should be as unified as possible. For example, when solid and outline figures occur together, the solid figure should be within the line outline figure, as shown in Figure 4-18e.

Standardization of Symbolic Displays

When symbol displays might be used in various circumstances by the same people, the displays should be standardized with a given symbol being associated with the same referent. One example of such standardization is the system of international road signs. A few examples are shown in Figure 4-19. The National Park Service also has a standardized set of symbols to represent various services and concepts such as picnic areas, bicycle trails, and playgrounds.

124

PART 2:

INFORMATION

INPUT

SLIPPERY ROAD

&
PEDESTRIAN CROSSING

6) cg
NO LEFT (RIGHT) TURN DON'T YOUR BLOW HORN

MEt:::HANICAL HEI..P

IT] W
rELEPHONE

(3) Danger SigM

Ib) Instruction signs

(c) Information

signs

FIGURE 4-19 Examples 01 a lew international countries. especially in Europe. erents.

road signs. These are standardized across many Most 01 these signs are directly symbolic 01 their rei

CODES

As discussed in Chapter 3, sometimes there is a need to have a coding system that identifies various items of a given class. For example. various specific pipes in a plumbing system can be coded different colors, or various types of highway signs can be coded by different shapes. The items to be coded (such as different pipes or different types of highway signs) are called . Thus, a specific code such as an octagonal (eight-sided) road sign means "stop." The types of visual stimuli used in such circumstances are sometimes called coding . Some visual coding dimensions are color, geometric and other shapes, letters. numerals, flash rate (of a flashing light), visual angle {the positions of hands on a clock), size (variation in the size of circles), and even "chart junk" [the term used by Tufte (1983) for the gaudy variations in shading and design of areas of graphs].
Single Coding Dimensions

When various coding dimensions can be used, an experiment can be carried out to determine what dimension would be best. One such study done by Smith and Thomas (1964) used the four sets of codes shown in Figure 4-20 in a task of counting the number of items of a specific class (such as red. gun, circle, or B-52) in a large display with many other items of the same class. (The density of items, that is. the total number of items in the display, was also varied.) The results, shown in Figure 4-21, show a clear superiority for the color codes for this task; this superiority was consistent for all levels of density. Although these and other studies indicate that various visual coding dimensions differ in their relevance for various tasks and in various situations, definite guidelines regarding the use of such codes still cannot be laid down. Recognizing this, and realizing that good judgment musteater into the selection of visual codes for specific purposes, we see that a comparison such as that given in Table 4-5 can serve as at least partial guidance. In particular, Table 4-5 indicates the approximate number of levels of each of various visual codes that can be discriminated, along with some sideline comments about certain methods.

CHAPTER

4:

TEXT. GRAPHICS.

SYMBOLS. AND CODES

125

~~~~~~ __-; .rC;' _


r"ongle! GMeometriC forms
---tt-

~.'S'eFm~lc;-,OrCOle_RFC+,r'C_OI

2 1,;_1
e Star

Dlom.~nJ

A!.:"
..

i
I'

!
Gun :
I! --

=-:i---~r-----I
Alrcrott Missile

I.

Radar ~ llitary

J
~-_.Green

Lymbol:'

--~--.--t-i --""'----t-.-8lue!
lS8G4/51 I

Ship -t-_-..IIoiMr Yellow


(10YR6/iO)

-1

II'~ (~~~o;:11
L~tatIOO)

White

Red
(SR4/9)

(2SG5/BI

(SyB/41

'stem ecific .es of .ch as 1US, a , The oding other : (the even ading

-}

FIGURE 4-20 Four sets of codes used in a study by Smith and Thomas. The notations under the color labels are the Munsell color matches of the colors used. (Source: Smith and Thomas, 1964. Copyright 1964 by the American Psychological Association and reproduced by permission.)

<II

-g40
0 U

80

'.; "
<II

u
+-

~30

g60

+c:

E 20
0

'"

'" s
<II

0'

'0

E 40
. Military syrroots '0
.\xl
+-

"

i2 10

'" ~20
o,

xl out hand isk of Ie, or sity of ) The es for irnenlions, Iown.


ection

+'
c
0

'"

'"

~
0 20
(0)

'" ~

60 Data on mean time

100

0
Items
(b)

Densrrv, nurnber of displayed

Octo on errors

FIGURE 4-21 Mean time a and errors b in counting items of four classes of codes as a function of display density. The Xs indicate comparison data for displays 01 100 items with color (or shape) held constant. (Source: Smith and Thomas, 1964. Copyright 1964 by the American Psychological Association and reprinted by permission.)

Color Coding

s that
ile 4-5

codes ertain

Since color is a fairly common visual code, we discuss it somewhat further. An important question relating to color deals with the number of distinct colors persons with normal color vision can differentiate on an absolute basis. It has generally been presumed that the number was relatively moderate; for example, Jones (1962) indicated that the normal observer could identify about nine

126

PART 2:

INFORMATION

INPUT

TABLE 4-5 SUMMARY OF CERTAIN VISUAL CODING METHODS (Numbers refer 10 number of levels which can be discriminated on an absolute basis under optimum conditions.) Alphanumeric Single numerals, 10; single letters, 26; combinations, unlimited. Good; especially useful for identification; uses little space if there is good contrast. Certain items easily confused with each other. Hues, 9; hue, saturation, and brightness combinations, 24 or more. Preferable limit, 9. Particularly good for searching and counting tasks. Affected by some lights; problem with color-defective individuals. *t 10. Preferable limit, 3. limited reading.t space required. Good for qualitative

Color (of surfaces)

Color (of lights) Geometric shapes

15 or more. Preferable limit,S. Generally useful coding system, particularly in symbolic representation; good for CRTs. Shapes used together need to be discriminable; some sets of shapes more difficult to discriminate than others.t 24. Preferable limit, 12. Generally satisfactory for special purposes such as indicating direction, angle, or position on round instruments like clocks, CRTs, etc. 5 or 6. Preferable limit, 3. Takes considerable space. Use only when specifically appropriate. 6. Preferable limit, 4. Use only when specifically appropriate, such as to represent numbers of items. Takes considerable space; may be confused with other symbols. 3-4. Preferable limit. 2. Use only when specifically appropriate. Weaker signals may be masked.t Preferable limit, 2. limited applicability if receiver needs to differentiate flash rates. Flashing lights, however, have possible use in combination with controlled time intervals (as with lighthouse signals and naval communications) or to attract attention to specific areas.

Angle of inclination

Size of forms (such as squares) Visual number

Brightness of lights Flash rate of lights

*Feallock et al. (1966). tM. R. Jones (1962). j:Grether and Baker (1972). Muller et al. (1955).

surface colors. varying primarily in hue. However. it appears that, with training. people can learn to make upward of a couple of dozen such discriminations when combinations of hue, saturation. and lightness are prepared in a nonredundant manner (Feallock et al., 1966). When relatively untrained people are to use color codes. however. the better part of wisdom would argue for the use of only a small number of discriminable colors. In general. color coding has been found to be particularly effective for "searching" tasks-tasks that involve scanning an array of many different stimuli to "search out" (or spot). or locale. or count those of a specific class. This advantage presumably is owing to ,the fact that colors "catch the eye" more readily than other visual codes. Examples of searching tasks include the use of maps and navigational charts. searching for items in a file. and searching

CHAPTER 4:

TEXT. GRAPHICS.

SYMBOLS.

AND CODES

127

for color-coded electric wires of some specific color. Christ (1975) analyzed the results of 42 studies in which color codes were compared with other types of codes and found that color codes were generally better for searching tasks than most other codes (geometric shapes. letters, and digits). Christ also found that color codes were better for' 'identification" tasks than were certain other codes, but color codes were generally not as good for such tasks as letter and numerals were. In this sense. an identification task involves the conceptual recognition of the meaningfulness of the code-not simple visual recognition of items of one color from others (such as identifying red labels-from all other labels). Thus, we can see the possible basis for the advantage of letters and numerals, since they could aid in tying down the meaning of codes (such as grade A or size 12). In general, then, color is a very useful coding dimension. but it is obviously not a universally useful system.
Multidimension Codes

Chapter 3 included a discussion of multidimension codes. indicating some of the variations in combining two or more dimensions. Heglin (1973/ recommends, however, that no more than two dimensions be used together if rapid interpretation is required. In combining two visual codes. certain combinations do not "go well" together. Some simply cannot be used in combination. The combinations that are potentially useful are shown as Xs in Figure 4-22. Although combinations of codes can be useful in many circumstances. they are not always more effective than single-dimension codes.

FIGURE 4-22 Potential combinations of coding systems for use in muJDdimension coding. (Source: Adapted from Heglin, 1973, Tables VI-6, VI22.)

~ ~
o
c
(;

c Q

.,
0

0
U

"
X

E
Z

'o, " , .c
0
!/l

~ ~ c '" .c
N

c
c Q (;

!/l

'"

ro
x x
X

"!'

-, x x

u 0

'~ " ~
U.

~ c ~ '" ~
0...J

'!! (;

2 ,
0'

<:

CoroNu"-e'ol Shape' S,ze B"ghtness


L OCOllOo DnO
'1'1

x
X

x x x
X

x I x

x
\

"
X

X X

X X
X

X X X X

X X

F losh rote L one length Angulor or .eniol .01"

X X X

..

128

PART 2:

INFORMA nON

INPUT

DISCUSSION

In our everyday lives. we are bombarded with scads of text, graphics. symbols, and codes. This chapter only scratched the surface of human factors issues involved with the display of such information. In the information age with increasing use of computers, more attention will be paid to human factors issues involved in the display of information in a form that can be quickly and accurately used without placing a burden on the receiver. Software packages such as word processors, data base programs, or file management utilities often fail or succeed in the marketplace based on how well information, both on the screen and in hardcopy manuals, is presented. Design of screen layouts and instruction manuals will continue to be areas in which human factors will make beneficial contributions.
REFERENCES Barfield, W., and Robless, R. (1989). The effects of two- or three-dimensional graphics on the problem-solving performance of experienced and novice ~~;on makers. 8, 369-385. Blakemore, C.; and Campbell. F. (1969). On the existence of neurones in the human visual system selectively sensitive to the orientation and size of retinal images. , 203, 237-260. perception Boff, K., Kaufman, L., and Thomas, J, (eds.) (1986). , vol. I: New York: Wiley. Booher, H. (1975). Relative comprehensibility of pictorial information and printed words in proceduralized instructions. 17,266-277. Broadbent, D. (1977). Language and ergonomics. 8, 15-18. Brown, C. (1988). design guidelines. Norwood, NJ: Ablex. Burg, A. (1966). Visual acuity as measured by static and dynamic tests: A comparative evaluation. 50,460-466. Cairney , P., and Siess, D. (1982). Communication effectiveness of symbolic safety signs with different user groups. 13,91-97. Christ, R. (1975). Review and analysis of color coding research for visual displays. , 17,542-570. Collins, B. and Lerner, N. (1983). (NBSIR 82-2685). Washington, DC: National Bureau of Standards. Cornog, D., and Rose, F. (1967, February). vol. 2: (NBS Misc. 262-2). Washington, DC: Government Printing Office. Cushman, W. (1986). Reading from microfiche, a VDT, and printed page and subjective fatigue. 28( I), 63-73. Easterby. R. (1967). Perceptual organization in static displays for man/machine systems. 10, 195-205. Easterby, R. (1970). The perception of symbols for machine displays. , 13, 149-158. Ellis. J . and Dewar, R. (1979). Rapid comprehension of verbal and symbolic traffic sign messages. 21, 161-168.

You might also like