- Email: [email protected]
Factors affecting preference ratings of prohibitive symbols
ARTICLE IN PRESS
Applied Ergonomics 34 (2003) 581–587
Factors affecting preference ratings of prohibitive symbols Kong-king Shieh*, Shih-miao Huang Department of Industrial Management, National Taiwan University of Science and Technology 43, Kee-Lung Road, Section 4, Taipei 106, Taiwan, Republic of China Received 7 September 2002; received in revised form 15 April 2003; accepted 1 June 2003
Abstract A sign consisting of a pictorial overlaid with a red circle-slash (i.e., a red circle with a red slash) is used ubiquitously to convey the message that some activity is prohibited. The purpose of this study was to investigate the effects of pictorial solidity, size, and direction of elongation (DE) of the pictorial and orientation and thickness of the red circle-slash on the preference ratings for prohibitive symbols. Solid (filled) pictorials were rated better than pictorials in outline form. Pictorials with a size equal to or greater than 75% of the length of the inner diameter of the circle-slash were rated higher than pictorials 50% in size. The effect of pictorial DE was not significant: pictorials with a greater vertical DE (i.e., tall/thin pictorials) did not differ from pictorials with a greater horizontal DE (i.e., short/wide pictorials), in terms of their preference ratings. However, pictorial DE interacted with slash orientation. Diagonal slashes were rated better than vertical or horizontal ones. Further, symbols were rated better when the thickness of the red circleslash was such that its resulting area comprised 25% of the total area inside its outer circle at least. Moreover, the interaction of pictorial size and slash thickness indicated that the preference for prohibitive symbols of thicker slash and smaller pictorial size might be degraded drastically. Implications of the results for the design of prohibitive symbols were discussed. r 2003 Elsevier Ltd. All rights reserved. Keywords: Prohibitive symbol design; Preference ratings; Solidity and size of pictorial; Orientation and thickness of slash
1. Introduction Warning symbols are extensively used as an interface by which critical situation-specific information is communicated to humans. There are two kinds of warning symbols: permissive and prohibitive. Permissive symbols refer to the symbols with positive messages providing information about permitted practices or encouraged behaviors, and prohibitive symbols refer to the symbols with negative messages frequently conveying information about actions that should not be taken in specific situations, or about conditions to be prevented or avoided (Glover et al., 1996). Gough (1965) believed that positively stated concepts are more easily understood than negative ones. Previous studies found that permissive symbols were better than prohibitive symbols (Ashley et al., 1971; Dewar and Swanson, 1972; Dewar, *Corresponding author. Tel.: +1-886-02-2737-6332; fax: +886-022737-6344. E-mail address: [email protected] (K. Shieh). 0003-6870/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0003-6870(03)00078-4
1976). Besides Ringseis and Caird (1995) believed that people might have difficulties in comprehension of negation when a warning slash was absent across some of the pictorials. However, Dewar (1976) thought that some concepts are difficult to express in a direct permissive way although permissive symbols may be better than prohibitive symbols with regard to glance legibility. Generally, a red circle-slash (i.e., a red circle with a red slash) is suggested to symbolize ‘‘prohibition’’ (ISO 3864, 1984; ANSI Z535.2, 1991). The meaning of a red circle with a slash is easily understood by most people because the concept has been applied so ubiquitously, e.g., with traffic signs. Thus, most people have learned that the red circle-slash is used to express the meaning, ‘‘no.’’ A prohibitive symbol normally includes a pictorial within an overlaid red circle-slash. The pictorial represents the item or activity that is prohibited. Difficulties attributable to the discrimination of the pictorial may increase when the aggregate symbol forms a ‘‘bad feature’’ (Garner, 1974), or when a distinctive
ARTICLE IN PRESS 582
K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
feature of the pictorial is obscure (Dewar, 1976; Biederman, 1987; Martindale, 1991; Murray et al., 1998). Symbols with ‘‘bad features’’ and/or ‘‘omitted distinctive features’’ could result from the improper design of either the slash or the pictorial. Thus, the discrimination of the prohibitive symbol may be affected by both the properties of the pictorial and the superimposed slash. Dewar (1976) assessed the effects of slash on glance legibility by examining four prohibition symbol variations: a red ring with a slash over the symbol, a red ring with a slash under the symbol, a red ring with a partial slash, and a slash only (not a prohibitive sign). He found that performance was best with no slash, and that a red ring with a slash over the symbol had worst performance. He reasoned that the conventional circle-slash declined performance because it would increase pictorial complexity and obscure portions of the pictorial. Hoffmann et al. (1982) measured two aspects of traffic sign legibility, distance legibility and glance legibility. They noted that circle-slashes under the pictorials were generally superior to circle-slashes over the pictorials for both measures. When circle-slashes were under pictorials, slashes with top right to bottom left were better than slashes from top left to bottom right. When circle-slashes were over pictorials, slashes with top right to bottom left were also better than slashes from top left to bottom right for distance legibility. However, slash orientation had no significant effect for glance legibility. In addition, Murray et al. (1998) investigated whether people’s judgments of four types of the circle-slash (a slash over the symbol, a slash under the symbol, a partial slash, and translucent slash) would differ in perceived effectiveness by preference rankings. They found that the over and under slashes were more preferable to translucent or partial slashes. Dewar’s (1976) finding that ‘‘symbol over slash’’ was worst was not found in the study. Wogalter et al. (2002) proposed that the inconsistent results may arise from the difference in dependent measures (preference rating vs. glance legibility) used in the two studies. They employed glance legibility to evaluate symbols used in Murray et al.’s experiment. They found that the subjects’ performance for under and translucent slashes were higher than for over or partial slashes. Recently, Shieh and Huang (in preparation for publication) investigated the effects of pictorial size and slash thickness on glance legibility under degraded situations. They found that the pictorial size may not be less than 75% and that slash thickness may not be thicker than 35% under degraded situations. However, those researches discussed only a small number of independent variables, such as slash type and slash orientation, although the design elements of a prohibitive symbol are numerous including the properties of
both pictorials and slashes. Therefore, it is needed to consider the effects of other design properties on prohibitive symbol design. Most pictorials in prohibitive symbols used for public information are either outline figure or solid figures. Solid pictorials are shapes that are filled. Outline pictorials are shapes with only a line in their boundaries. Solid pictorials appear more conspicuous than outline pictorials, and solid shapes are more legible than the outline shapes. Studies by Easterby (1970) and Sanders and McCormick (1993) found that solid figures were clearly superior to outline figures although not using prohibitive symbols. Increasing the size of pictorials improves legibility. However, beyond a certain size legibility attains asymptotic levels and may even deteriorate (Bullimore et al., 1991). Pictorial size may determine the degree to which important details of prohibitive symbols are concealed by the slash. A thick slash would conceal a greater portion of a small pictorial than would a large pictorial, and pictorial recognition could be more difficult. Shieh and Huang (in preparation for publication) found that pictorial size should be at least 75% of inner diameter of circle-slash. How pictorial size interacts with other symbol design variables, such as slash orientation and thickness, deserves further study. The position of the slash within the circle may determine the extent to which important features of a pictorial are concealed by the slash. The ANSI Z535.2 (1991) and ISO 3864 (1984) suggested that the slash be placed at a 45 diagonal (top left to bottom right) of the circle. However, the degree to which critical features of a pictorial are hidden might be determined by the slash orientation and the pictorial’s greatest direction of elongation (DE in short). Thus, a horizontal slash might hide fewer critical features than a 45 slash when the pictorial’s greatest DE is vertical. A tall/thin pictorial would have a greater vertical DE; a short/wide pictorial, a greater horizontal DE. A need exists to determine the effect of pictorial DE and slash orientation on the legibility of prohibitive symbols. The thickness of the red circle-slash is another factor that may affect the degree to which important features are concealed by the slash. The ISO 3864 (1984) recommended that the area of the red circle-slash included in the warning symbol be at least 35% of the total area inside the outer circle, leaving 65% of the area for the pictorial. Generally, a thicker slash would cover more details of the pictorial, thus increasing the difficulty in understanding the pictorial. Further, the slash separates the pictorial into two parts. Thicker slashes produce more separation between the two pictorial parts and further degrade the completeness of the pictorial. When the completeness of the pictorial is degraded, the prohibitive symbol may not be perceived as a ‘‘good shape,’’ and the difficulty in recognizing the
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
pictorial would increase. Moreover, there is scant literature or empirical research findings to support the ISO recommendation that the red circle-slash should be at least 35% of the total area of the prohibitive symbol. The main purpose of the present study was to investigate characteristics of two components of prohibitive symbols and their effects on subjects’ preference. One component was the pictorial, and the characteristics examined were its pictorial solidity, size and DE. The other component was the circle-slash; its slash orientation and thickness were investigated.
583
was defined as per ISO 3864 (1984), i.e., (area of the redcircle slash/total area inside the outer circle of the prohibitive symbol) 100%. Fig. 1 shows the prohibitive symbols used in the present study with their design descriptions. 2.2. Subjects
2. Method
The subjects were 14 male and 10 female undergraduate students from Oriental Institute of Technology. They were paid $ 10US for their participation in the experiment. Their mean age was 22.63 years with a standard deviation of 2.90 years. Twelve subjects were from the Industrial Design Department; the other twelve were from other disciplines.
2.1. Experimental design
2.3. Materials
The experiment was a 2 (pictorial solidity) 3 (pictorial size) 2 (pictorial DE) 4 (orientation of the red circle-slash) 3 (thickness of the red circle-slash) within-subjects split-plot design. Pictorial solidity had 2 levels: outline and solid. Pictorial size was defined as (the length of the pictorial/diameter of inner circle of prohibitive symbol) 100%. Size had three levels: 50%, 75%, and100%. There were two levels of pictorial DE: vertical (i.e., the pictorial had a greater vertical dimension) and horizontal (i.e., the pictorial had a greater horizontal dimension). Slash orientation had four levels: 0 (horizontal), 45 (diagonal from top left to bottom right), 90 (vertical) and 135 (diagonal from top right to bottom left). The thickness of the red circleslash had three levels: 25%, 35%, and 45%. Thickness
Six prohibitive symbols (see Fig. 1) were selected from traffic signs and warning symbols used in Taiwan. The pictorials in three symbols had a greater vertical DE; whereas, in the other three symbols the pictorials had a greater horizontal DE. For each symbol, 72 designs were constructed according to the combination of pictorial solidity (two levels) and size (three levels) of the pictorials, and the orientation (four levels) and thickness (three levels) of the red circle-slash. The prohibitive symbols, 8 cm in diameter, were drawn in the center of 10 10 cm cards. Each symbol was composed of a pictorial and a red circle-slash superimposed and centered on the pictorial (ISO 3864, 1984).
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 1. The six prohibitive symbols used in the experiment. Meaning and design characteristics of the symbols are described below. Design characteristics are expressed in the order of: sign meaning, pictorial solidity, pictorial size, pictorial DE, slash orientation and slash thickness. (A) No pedestrians, outline, 100%, vertical, 0 and 25%. (B) No swimming, outline, 75%, horizontal, 135 and 25%. (C) No touching, outline, 50%, vertical, 90 and 45%. (D) No bus entrance, solid, 100%, horizontal, 135 and 45%. (E) No car entrance, solid, 75%, horizontal, 45 and 35%. (F) No locking, solid, 50%, vertical, 0 and 35%.
ARTICLE IN PRESS 584
K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
2.4. Procedure At the beginning, the subjects were told the meanings of the prohibitive symbols. For each symbol, the cards were presented to the subjects in a random order. Also the order in which the six prohibitive symbols were presented was random across the subjects. The subjects were instructed to rate the preference of the 72 cards for each prohibitive symbol. They were told that the study was to investigate the identification of pictorials presented when they were covered by a red circle-slash. The instructions emphasized that evaluation of each symbol should consider potential environments, such as rain, fog, etc. Then, each subject sorted the cards into 10 groups based on their preference, with the constraint that at least three but not more than 12 cards are assigned to each group. After the subjects sorted the 72 cards for each symbol, the experimenter scored the cards. The cards in the least preferred group were scored ‘‘1’’; the next preferred group, ‘‘2,’’ and so forth with the most preferred group being scored ‘‘10.’’ This procedure continued until all six sets of cards were scored. The experimental task, per subject, lasted about 1 h. 2.5. Data analysis Analysis of variance and regression analysis were employed to analyze the data. All calculations were made using the Statistical Analysis System (SAS).
3. Results Table 1 shows the means and standard deviations of preference ratings for each level of the independent variables. Results of analysis of variance indicated that the main effect of pictorial solidity was significant (F (1, 23)=18.85, po0:01). Solid pictorials (5.68) were rated better than outline pictorials (5.23). There was also a significant main effect (F (2, 46)=128.04, po0:01) for pictorial size. Multiple comparisons using the Duncan’s method showed that preference ratings for the 100% pictorial size (6.49) and 75% size (6.14) did not significantly differ, but ratings for the 50% size (3.73) were significantly lower. The main effect of pictorial DE was not statistically significant. As for the effects of the slash properties, the results showed a significant main effect of slash orientation (F (3, 69)=153.37, po0:01) on preference ratings. Duncan’s multiple range test showed that the difference of ratings for the 45 (7.08) and 135 (7.01) orientations was not significant, nor was the difference between the 90 (3.69) and 0 (4.03) orientations. However, the ratings for the former two orientations were significantly greater than those of the latter. There was also a
Table 1 Means and standard deviations of preference ratings for each level of the independent variables Variables
Level
Mean (SD)
Duncan group
Pictorial solidity
Solid Outline
5.68 (3.15) 5.23 (3.10)
A B
Pictorial size
50% 75% 100%
3.73 (2.98) 6.14 (3.01) 6.49 (2.64)
B A A
Pictorial direction of elongation (DE)
Horizontal Vertical
5.50 (3.26) 5.41 (3.00)
A A
Red circle-slash thickness
25% 35% 45%
6.05 (3.16) 5.56 (2.97) 4.75 (3.13)
A B C
Slash orientation
0 45 90 135
4.03 7.08 3.69 7.01
A B A B
(2.42) (3.24) (2.27) (2.76)
Different letters in Duncan group indicate significant differences at 0.01 level.
significant main effect (F (2, 46)=40.37, po0:01) for thickness of the red circle-slash. Multiple comparisons using the Duncan method showed that the 25% slash thickness was rated most preferred (6.05), with 35% thickness next (5.56), and 45% thickness the least preferred (4.75); all differences were significant. Table 2 and Fig. 2 show preference ratings under combinations of slash orientation and pictorial DE. The interaction of pictorial DE and slash orientation was significant (F (3, 69)=72.86, po0:01). Analysis of the simple main effect indicated no significant differences between the vertical and horizontal pictorials when the slash orientation was 45 or 135 . However, the differences between the two pictorial directions of elongation were significant when the slash orientation was 0 or 90 . When the slash was horizontal (0 ), the ratings for vertical pictorials were better than those for horizontal; however, the horizontal pictorials were rated better than the vertical ones when the slash was vertical (90 ). There was also a significant interaction (F (4, 92)=21.03, po0:01) between pictorial size and slash thickness as shown in Table 3 and Fig. 3. Analysis of simple main effects indicated that when the pictorial size was 75% or 100%, the ratings for 25% slash thickness and 35% thickness were not significantly different, but that for 45% slash thickness was significantly lower. When the pictorial size was 50%, rating for 25% slash thickness was the best, with that of 35% slash thickness the next, and 45% slash thickness the lowest.
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
585
Table 2 Preference rating scores under combinations of slash orientation and pictorial direction of elongation Slash orientation Variables
Level
Pictorial Direction of Elongation (DE)
0
Horizontal Vertical
90
45
135
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
3.23 (2.32) 4.84 (2.24)
A
6.85 (2.87) 7.16 (2.64)
A
4.79 (2.11) 2.59 (1.85)
A
7.11 (3.79) 7.05 (2.56)
A
B
A
B
Note: Different letters in Duncan group indicate significant differences at 0.01 level.
8
Preference Ratio
7 6
Slash Orientation
5
0 degree 45 degree 90 degree 135 degree
4 3 2 1 0 Horizontal (DE)
Vertical (DE)
Pictorial Direction of Elongation
Fig. 2. Interaction between slash orientation and pictorial DE.
4. Discussion The major purpose of this study was to investigate the effects of pictorial solidity, size and DE, and orientation and thickness of the red-circle slash on preference ratings for prohibitive symbols. Data analysis showed that solid pictorials were rated more preferable than outline pictorials. This result is consistent with previous research findings (cf. Sanders and McCormick, 1993). A plausible explanation is that outline figures are visually more complicated because they are constructed with more lines, angles and segments than solid figures (Hochberg and Brooks, 1960). Many studies found that complicated figures degraded human performance (Easterby and Zwaga, 1984; Curry et al., 1998; Dewar, 1999). For pictorial size, preference ratings increased as size increased from 50% to 100%. However, there was no significant improvement in ratings from 75% to 100% pictorial size. Generally speaking, larger pictorials are easier to recognize, and hence would receive better preference. Caird et al. (1997) found that small pictorials and pictorials with slashes result in difficulties in comprehension. This finding also agreed with Matlin’s finding (1989) that when the area hidden by a slash did not exceed a certain amount, symbol recognition was not degraded. Based on the results of this study, a pictorial size that is 75% of the inner diameter of the red circle-slash is appropriate, and 100% is not necessary.
Shieh and Huang (in preparation for publication) had similar finding. They investigated the effects of pictorial size and slash thickness on glance legibility of prohibitive symbols. Their results showed that a pictorial size that is 75% of the inner diameter of the red circle-slash is appropriate under degraded environment. For pictorial DE, the preference ratings for vertical and horizontal pictorial were similar. Pictorials with a greater vertical DE (i.e., tall/thin pictorials) did not significantly differ from pictorials with a greater horizontal DE (i.e., short/wide pictorials) in terms of their preference ratings. However, the significant interaction between the pictorial’s DE and slash orientation indicated that preference rating was lower when the pictorial’s DE and slash orientation were the same. Possibly, this result occurred because when the pictorial’s DE and slash orientation were the same, the slash would obscure more features of the pictorial and perhaps make it harder to recognize. By the same token, a diagonal slash (45 or 135 ) would obscure less of the pictorial than would a slash aligned with the pictorial’s greatest DE and thus be rated more preferable, as found in this study. However, the greater preference ratings for diagonal slashes, even where vertical slashes overlaid horizontal pictorials, or horizontal slashes overlaid vertical pictorials, indicated that other factors might also be involved. For instance, subjects might have been familiar with the existing prohibitive symbols with diagonal slashes, which they see in their daily activities and thus rated the diagonal slashes higher. Thickness of the red circle-slash significantly affected preference ratings. Subjects rated the 25% thickness the best, 35% next, and 45% the worst. Preference ratings decreased as thickness of the red circle-slash increased. Besides, the significant interaction between pictorial size and slash thickness indicated that preference rating increase for 75% or 100% pictorial size leveled off when the thickness was below 35%. However, subjects rated the 25% thickness the best, 35% next, and 45% the worst when the pictorial size was 50%. The evidence indicated that the preference for prohibitive symbols of thicker slash and smaller pictorial size might be degraded
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
586
Table 3 Preference rating scores under combinations of pictorial size and slash thickness Pictorial size Variables
Level
Slash thickness
25% 35% A 45%
50%
75%
100%
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
4.68 (3.66) 3.78 (2.57) 2.74 (2.21)
A
6.77 (2.71) 6.25 (2.87) 5.39 (3.26)
A
6.70 (2.54) 6.65 (2.61) 6.11 (2.75)
A
B C
A B
A B
Note: Different letters in Duncan group indicate significant differences at 0.01 level.
Preference Ratio
8 7 6 5 4 3
Slash Thickness 25% 35%
2
45%
1 0 50%
75%
100%
Pictorial Size
Fig. 3. Interaction between pictorial size and slash thickness.
drastically. Apparently, the greater the thickness, the lesser the available space for the pictorial and the more the pictorial was covered by the slash, and hence resulted in lower ratings, especially when pictorial size was small. Therefore, to acquire better symbol preference, the slash thickness may be thin enough when the pictorial size is small. However, Subjects may overlook the red slash shape in prohibitive symbol when the slash thickness is not thick enough. In order to keep clear legibility of both pictorials and slashes in prohibitive symbols, the slash size should not be too small. Based on this result, the recommendation of ISO 3864 (1984) that the red circleslash should be at least 35% of the area of the sign may be suitable when the pictorial size is not too small. But, it still needs further study. For slash orientation, ratings for the diagonal (45 and 135 ) slashes were similar, and were rated higher than those for the non-diagonal (0 and 90 ) slashes. This may give the designer of prohibitive symbols some flexibility to use a 45 or 135 slash depending on which would obscure the pictorial to a lesser degree. These results also agreed with Hoffmann et al.’s (1982) findings that slashes with top right to bottom left and slashes from top left to bottom right had no significant differences for glance legibility when circleslashes were over pictorials. However, they also found that slashes with top right to bottom left were better than slashes from top left to bottom right for distance legibility. Therefore, more research is needed to replicate these results.
Finally, the features of shapes are various. Beside vertical DE, or horizontal DE, there are square shape, L shape, T shape, O shape, etc. The six pictorials used in present study cannot represent all the other kinds of pictorial. Whether the results of this study are also valid for other types of pictorials requires further study. The results of this study were very encouraging. Variables that significantly affect the preference rating of prohibitive symbols were identified and their relative effects quantified. Useful follow-on efforts include studies that utilize objective recognition performance measures in tasks that are more closely appropriate realworld tasks, and compare that performance with the predicted preference rating based on the results from the present study.
5. Conclusion Based on the results of the present study, the following conclusions are made concerning the design of prohibitive symbols with respect to preference rating: (1) Solid pictorials are better than outline pictorial. (2) Pictorial size should be at least 75% of the diameter of the inside circle of the prohibitive symbol. (3) The pictorial’s greatest DE and the slash orientation should not be the same. In other words, short/ wide pictorials should not be used with horizontal slashes; nor should tall/thin pictorials be used with vertical slashes. (4) Diagonal slashes (45 or 135 ) are preferred, versus, horizontal or vertical slashes. (5) The thickness of the red circle-slash may be as small as 35% as suggested by ISO 3864 (1984).
Acknowledgements This study was supported by a Research Grant from the National Sciences Council of the Republic of China, Grant No. NSC-89-2213-E-161-011.
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
References ANSI, 1991. American National Standard for Environmental and Facility Safety Signs: Z535.2. American National Standard Institute, Washington, DC. Ashley, F., Myer, J., Smith, D., 1971. City signs and lights, Boston Redevelopment Authority, January. Biederman, I., 1987. Recognition by components. Psychol. Rev. 4, 115–147. Bullimore, M.A., Fulton, E.J., Howarth, P.A., 1991. Assessment of visual performance. In: Wilson, J.R. and Corlett, E. (Eds.), Evaluation of Human Work. Taylor & Francis, London, pp. 648–681 (Chapter 26). Caird, J.F., Wheat, B., McIntosh, K.R., Dewar, R.E., 1997. The comprehensibility of airline safety card pictorials. In: Proceedings of the Human Factors and Ergonomics Society 41st Annual Meeting, Santa Monica, pp. 801–805. Curry, M.B., McDougall, S.J.P., Bruijn, O. 1998. The effects of the visual metaphor in determining icon efficacy. In: Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting, Santa Monica, CA, pp. 1590–1594. Dewar, R., 1999. Design and evaluation of public information symbols, In: Zwaga, H.J.G. (Ed.), Visual Information for Everyday Use. Taylor & Francis, London, pp. 285–303. Dewar, R.E., 1976. The slash obscures the symbol on prohibitive traffic signs. Hum. Factors 18 (3), 253–258. Dewar, R.E., Swanson, H.A., 1972. Recognition of traffic control signs. Highw. Res. Rec. 414, 16–23. Easterby, R., 1970. The perception of symbols for machine displays. Ergonomics 13, 145–158. Easterby, R., Zwaga, H., 1984. Information Design: The Design and Evaluation of Signs and Printed Material. Wiley, New York, 588pp.
587
Garner, W.R., 1974. The Processing of Information and Structure. Erlbaum, Potomac, MD. Glover, B.L., Magurno, A.B., Murray, L.A., Wogalter, M.S., 1996. Pictorial Negations: preferences for different circle-slash variations. In: Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting, Santa Monica, pp. 910–914. Gough, P.B., 1965. Grammatical transformation and speed of understanding. J. Verbal Learn. Verbal Behav. 4, 107–111. Hochberg, J., Brooks, V., 1960. The psychophysics of form: reversibleperspective drawings of spatial objects. Am. J. Psychol. 73, 337–354. Hoffmann, E.R., Smith, G., Kalnins, C., Sanders, B., 1982. Legibility of symbolic turn-restriction signs of various formats. Proc. Aust. Road Res. Board 11 (5), 127–134. ISO 3864, 1984. International Standard for Safety Colours and Safety Signs. ISO, Switzerland. Martindale, C., 1991. Cognitive Psychology: A Neural-Network Approach. Cole, CA. Matlin, M.W., 1989. Cognition. Holt, Rinehart and Winston, NY. Murray, L.A., Magurno, A.B., Glover, B.L., Wogalter, M.S., 1998. Prohibitive pictorials: evaluations of different circle-slash negation symbols. Int. J. Ind. Ergonomics 22, 473–482. Ringseis, E.L., Caird, J.K., 1995. The comprehensibility and legibility of twenty pharmaceutical warning pictogram. In: Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting, Santa Monica, pp. 974–978. Sanders, M.S., McCormick, E.J., 1993. Human Factors in Engineering Design. McGraw-Hill, New York, pp. 121–125. Wogalter, M.S., Murray, L.A., Glover, B.L., Shaver, E.F., 2002. Comprehension of different types of prohibitive safety symbols with glance exposure. In: Proceedings of the Human Factors and Ergonomics Society 46th Annual Meeting, Santa Monica, pp. 1753–1757.
Applied Ergonomics 34 (2003) 581–587
Factors affecting preference ratings of prohibitive symbols Kong-king Shieh*, Shih-miao Huang Department of Industrial Management, National Taiwan University of Science and Technology 43, Kee-Lung Road, Section 4, Taipei 106, Taiwan, Republic of China Received 7 September 2002; received in revised form 15 April 2003; accepted 1 June 2003
Abstract A sign consisting of a pictorial overlaid with a red circle-slash (i.e., a red circle with a red slash) is used ubiquitously to convey the message that some activity is prohibited. The purpose of this study was to investigate the effects of pictorial solidity, size, and direction of elongation (DE) of the pictorial and orientation and thickness of the red circle-slash on the preference ratings for prohibitive symbols. Solid (filled) pictorials were rated better than pictorials in outline form. Pictorials with a size equal to or greater than 75% of the length of the inner diameter of the circle-slash were rated higher than pictorials 50% in size. The effect of pictorial DE was not significant: pictorials with a greater vertical DE (i.e., tall/thin pictorials) did not differ from pictorials with a greater horizontal DE (i.e., short/wide pictorials), in terms of their preference ratings. However, pictorial DE interacted with slash orientation. Diagonal slashes were rated better than vertical or horizontal ones. Further, symbols were rated better when the thickness of the red circleslash was such that its resulting area comprised 25% of the total area inside its outer circle at least. Moreover, the interaction of pictorial size and slash thickness indicated that the preference for prohibitive symbols of thicker slash and smaller pictorial size might be degraded drastically. Implications of the results for the design of prohibitive symbols were discussed. r 2003 Elsevier Ltd. All rights reserved. Keywords: Prohibitive symbol design; Preference ratings; Solidity and size of pictorial; Orientation and thickness of slash
1. Introduction Warning symbols are extensively used as an interface by which critical situation-specific information is communicated to humans. There are two kinds of warning symbols: permissive and prohibitive. Permissive symbols refer to the symbols with positive messages providing information about permitted practices or encouraged behaviors, and prohibitive symbols refer to the symbols with negative messages frequently conveying information about actions that should not be taken in specific situations, or about conditions to be prevented or avoided (Glover et al., 1996). Gough (1965) believed that positively stated concepts are more easily understood than negative ones. Previous studies found that permissive symbols were better than prohibitive symbols (Ashley et al., 1971; Dewar and Swanson, 1972; Dewar, *Corresponding author. Tel.: +1-886-02-2737-6332; fax: +886-022737-6344. E-mail address: [email protected] (K. Shieh). 0003-6870/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0003-6870(03)00078-4
1976). Besides Ringseis and Caird (1995) believed that people might have difficulties in comprehension of negation when a warning slash was absent across some of the pictorials. However, Dewar (1976) thought that some concepts are difficult to express in a direct permissive way although permissive symbols may be better than prohibitive symbols with regard to glance legibility. Generally, a red circle-slash (i.e., a red circle with a red slash) is suggested to symbolize ‘‘prohibition’’ (ISO 3864, 1984; ANSI Z535.2, 1991). The meaning of a red circle with a slash is easily understood by most people because the concept has been applied so ubiquitously, e.g., with traffic signs. Thus, most people have learned that the red circle-slash is used to express the meaning, ‘‘no.’’ A prohibitive symbol normally includes a pictorial within an overlaid red circle-slash. The pictorial represents the item or activity that is prohibited. Difficulties attributable to the discrimination of the pictorial may increase when the aggregate symbol forms a ‘‘bad feature’’ (Garner, 1974), or when a distinctive
ARTICLE IN PRESS 582
K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
feature of the pictorial is obscure (Dewar, 1976; Biederman, 1987; Martindale, 1991; Murray et al., 1998). Symbols with ‘‘bad features’’ and/or ‘‘omitted distinctive features’’ could result from the improper design of either the slash or the pictorial. Thus, the discrimination of the prohibitive symbol may be affected by both the properties of the pictorial and the superimposed slash. Dewar (1976) assessed the effects of slash on glance legibility by examining four prohibition symbol variations: a red ring with a slash over the symbol, a red ring with a slash under the symbol, a red ring with a partial slash, and a slash only (not a prohibitive sign). He found that performance was best with no slash, and that a red ring with a slash over the symbol had worst performance. He reasoned that the conventional circle-slash declined performance because it would increase pictorial complexity and obscure portions of the pictorial. Hoffmann et al. (1982) measured two aspects of traffic sign legibility, distance legibility and glance legibility. They noted that circle-slashes under the pictorials were generally superior to circle-slashes over the pictorials for both measures. When circle-slashes were under pictorials, slashes with top right to bottom left were better than slashes from top left to bottom right. When circle-slashes were over pictorials, slashes with top right to bottom left were also better than slashes from top left to bottom right for distance legibility. However, slash orientation had no significant effect for glance legibility. In addition, Murray et al. (1998) investigated whether people’s judgments of four types of the circle-slash (a slash over the symbol, a slash under the symbol, a partial slash, and translucent slash) would differ in perceived effectiveness by preference rankings. They found that the over and under slashes were more preferable to translucent or partial slashes. Dewar’s (1976) finding that ‘‘symbol over slash’’ was worst was not found in the study. Wogalter et al. (2002) proposed that the inconsistent results may arise from the difference in dependent measures (preference rating vs. glance legibility) used in the two studies. They employed glance legibility to evaluate symbols used in Murray et al.’s experiment. They found that the subjects’ performance for under and translucent slashes were higher than for over or partial slashes. Recently, Shieh and Huang (in preparation for publication) investigated the effects of pictorial size and slash thickness on glance legibility under degraded situations. They found that the pictorial size may not be less than 75% and that slash thickness may not be thicker than 35% under degraded situations. However, those researches discussed only a small number of independent variables, such as slash type and slash orientation, although the design elements of a prohibitive symbol are numerous including the properties of
both pictorials and slashes. Therefore, it is needed to consider the effects of other design properties on prohibitive symbol design. Most pictorials in prohibitive symbols used for public information are either outline figure or solid figures. Solid pictorials are shapes that are filled. Outline pictorials are shapes with only a line in their boundaries. Solid pictorials appear more conspicuous than outline pictorials, and solid shapes are more legible than the outline shapes. Studies by Easterby (1970) and Sanders and McCormick (1993) found that solid figures were clearly superior to outline figures although not using prohibitive symbols. Increasing the size of pictorials improves legibility. However, beyond a certain size legibility attains asymptotic levels and may even deteriorate (Bullimore et al., 1991). Pictorial size may determine the degree to which important details of prohibitive symbols are concealed by the slash. A thick slash would conceal a greater portion of a small pictorial than would a large pictorial, and pictorial recognition could be more difficult. Shieh and Huang (in preparation for publication) found that pictorial size should be at least 75% of inner diameter of circle-slash. How pictorial size interacts with other symbol design variables, such as slash orientation and thickness, deserves further study. The position of the slash within the circle may determine the extent to which important features of a pictorial are concealed by the slash. The ANSI Z535.2 (1991) and ISO 3864 (1984) suggested that the slash be placed at a 45 diagonal (top left to bottom right) of the circle. However, the degree to which critical features of a pictorial are hidden might be determined by the slash orientation and the pictorial’s greatest direction of elongation (DE in short). Thus, a horizontal slash might hide fewer critical features than a 45 slash when the pictorial’s greatest DE is vertical. A tall/thin pictorial would have a greater vertical DE; a short/wide pictorial, a greater horizontal DE. A need exists to determine the effect of pictorial DE and slash orientation on the legibility of prohibitive symbols. The thickness of the red circle-slash is another factor that may affect the degree to which important features are concealed by the slash. The ISO 3864 (1984) recommended that the area of the red circle-slash included in the warning symbol be at least 35% of the total area inside the outer circle, leaving 65% of the area for the pictorial. Generally, a thicker slash would cover more details of the pictorial, thus increasing the difficulty in understanding the pictorial. Further, the slash separates the pictorial into two parts. Thicker slashes produce more separation between the two pictorial parts and further degrade the completeness of the pictorial. When the completeness of the pictorial is degraded, the prohibitive symbol may not be perceived as a ‘‘good shape,’’ and the difficulty in recognizing the
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
pictorial would increase. Moreover, there is scant literature or empirical research findings to support the ISO recommendation that the red circle-slash should be at least 35% of the total area of the prohibitive symbol. The main purpose of the present study was to investigate characteristics of two components of prohibitive symbols and their effects on subjects’ preference. One component was the pictorial, and the characteristics examined were its pictorial solidity, size and DE. The other component was the circle-slash; its slash orientation and thickness were investigated.
583
was defined as per ISO 3864 (1984), i.e., (area of the redcircle slash/total area inside the outer circle of the prohibitive symbol) 100%. Fig. 1 shows the prohibitive symbols used in the present study with their design descriptions. 2.2. Subjects
2. Method
The subjects were 14 male and 10 female undergraduate students from Oriental Institute of Technology. They were paid $ 10US for their participation in the experiment. Their mean age was 22.63 years with a standard deviation of 2.90 years. Twelve subjects were from the Industrial Design Department; the other twelve were from other disciplines.
2.1. Experimental design
2.3. Materials
The experiment was a 2 (pictorial solidity) 3 (pictorial size) 2 (pictorial DE) 4 (orientation of the red circle-slash) 3 (thickness of the red circle-slash) within-subjects split-plot design. Pictorial solidity had 2 levels: outline and solid. Pictorial size was defined as (the length of the pictorial/diameter of inner circle of prohibitive symbol) 100%. Size had three levels: 50%, 75%, and100%. There were two levels of pictorial DE: vertical (i.e., the pictorial had a greater vertical dimension) and horizontal (i.e., the pictorial had a greater horizontal dimension). Slash orientation had four levels: 0 (horizontal), 45 (diagonal from top left to bottom right), 90 (vertical) and 135 (diagonal from top right to bottom left). The thickness of the red circleslash had three levels: 25%, 35%, and 45%. Thickness
Six prohibitive symbols (see Fig. 1) were selected from traffic signs and warning symbols used in Taiwan. The pictorials in three symbols had a greater vertical DE; whereas, in the other three symbols the pictorials had a greater horizontal DE. For each symbol, 72 designs were constructed according to the combination of pictorial solidity (two levels) and size (three levels) of the pictorials, and the orientation (four levels) and thickness (three levels) of the red circle-slash. The prohibitive symbols, 8 cm in diameter, were drawn in the center of 10 10 cm cards. Each symbol was composed of a pictorial and a red circle-slash superimposed and centered on the pictorial (ISO 3864, 1984).
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 1. The six prohibitive symbols used in the experiment. Meaning and design characteristics of the symbols are described below. Design characteristics are expressed in the order of: sign meaning, pictorial solidity, pictorial size, pictorial DE, slash orientation and slash thickness. (A) No pedestrians, outline, 100%, vertical, 0 and 25%. (B) No swimming, outline, 75%, horizontal, 135 and 25%. (C) No touching, outline, 50%, vertical, 90 and 45%. (D) No bus entrance, solid, 100%, horizontal, 135 and 45%. (E) No car entrance, solid, 75%, horizontal, 45 and 35%. (F) No locking, solid, 50%, vertical, 0 and 35%.
ARTICLE IN PRESS 584
K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
2.4. Procedure At the beginning, the subjects were told the meanings of the prohibitive symbols. For each symbol, the cards were presented to the subjects in a random order. Also the order in which the six prohibitive symbols were presented was random across the subjects. The subjects were instructed to rate the preference of the 72 cards for each prohibitive symbol. They were told that the study was to investigate the identification of pictorials presented when they were covered by a red circle-slash. The instructions emphasized that evaluation of each symbol should consider potential environments, such as rain, fog, etc. Then, each subject sorted the cards into 10 groups based on their preference, with the constraint that at least three but not more than 12 cards are assigned to each group. After the subjects sorted the 72 cards for each symbol, the experimenter scored the cards. The cards in the least preferred group were scored ‘‘1’’; the next preferred group, ‘‘2,’’ and so forth with the most preferred group being scored ‘‘10.’’ This procedure continued until all six sets of cards were scored. The experimental task, per subject, lasted about 1 h. 2.5. Data analysis Analysis of variance and regression analysis were employed to analyze the data. All calculations were made using the Statistical Analysis System (SAS).
3. Results Table 1 shows the means and standard deviations of preference ratings for each level of the independent variables. Results of analysis of variance indicated that the main effect of pictorial solidity was significant (F (1, 23)=18.85, po0:01). Solid pictorials (5.68) were rated better than outline pictorials (5.23). There was also a significant main effect (F (2, 46)=128.04, po0:01) for pictorial size. Multiple comparisons using the Duncan’s method showed that preference ratings for the 100% pictorial size (6.49) and 75% size (6.14) did not significantly differ, but ratings for the 50% size (3.73) were significantly lower. The main effect of pictorial DE was not statistically significant. As for the effects of the slash properties, the results showed a significant main effect of slash orientation (F (3, 69)=153.37, po0:01) on preference ratings. Duncan’s multiple range test showed that the difference of ratings for the 45 (7.08) and 135 (7.01) orientations was not significant, nor was the difference between the 90 (3.69) and 0 (4.03) orientations. However, the ratings for the former two orientations were significantly greater than those of the latter. There was also a
Table 1 Means and standard deviations of preference ratings for each level of the independent variables Variables
Level
Mean (SD)
Duncan group
Pictorial solidity
Solid Outline
5.68 (3.15) 5.23 (3.10)
A B
Pictorial size
50% 75% 100%
3.73 (2.98) 6.14 (3.01) 6.49 (2.64)
B A A
Pictorial direction of elongation (DE)
Horizontal Vertical
5.50 (3.26) 5.41 (3.00)
A A
Red circle-slash thickness
25% 35% 45%
6.05 (3.16) 5.56 (2.97) 4.75 (3.13)
A B C
Slash orientation
0 45 90 135
4.03 7.08 3.69 7.01
A B A B
(2.42) (3.24) (2.27) (2.76)
Different letters in Duncan group indicate significant differences at 0.01 level.
significant main effect (F (2, 46)=40.37, po0:01) for thickness of the red circle-slash. Multiple comparisons using the Duncan method showed that the 25% slash thickness was rated most preferred (6.05), with 35% thickness next (5.56), and 45% thickness the least preferred (4.75); all differences were significant. Table 2 and Fig. 2 show preference ratings under combinations of slash orientation and pictorial DE. The interaction of pictorial DE and slash orientation was significant (F (3, 69)=72.86, po0:01). Analysis of the simple main effect indicated no significant differences between the vertical and horizontal pictorials when the slash orientation was 45 or 135 . However, the differences between the two pictorial directions of elongation were significant when the slash orientation was 0 or 90 . When the slash was horizontal (0 ), the ratings for vertical pictorials were better than those for horizontal; however, the horizontal pictorials were rated better than the vertical ones when the slash was vertical (90 ). There was also a significant interaction (F (4, 92)=21.03, po0:01) between pictorial size and slash thickness as shown in Table 3 and Fig. 3. Analysis of simple main effects indicated that when the pictorial size was 75% or 100%, the ratings for 25% slash thickness and 35% thickness were not significantly different, but that for 45% slash thickness was significantly lower. When the pictorial size was 50%, rating for 25% slash thickness was the best, with that of 35% slash thickness the next, and 45% slash thickness the lowest.
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
585
Table 2 Preference rating scores under combinations of slash orientation and pictorial direction of elongation Slash orientation Variables
Level
Pictorial Direction of Elongation (DE)
0
Horizontal Vertical
90
45
135
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
3.23 (2.32) 4.84 (2.24)
A
6.85 (2.87) 7.16 (2.64)
A
4.79 (2.11) 2.59 (1.85)
A
7.11 (3.79) 7.05 (2.56)
A
B
A
B
Note: Different letters in Duncan group indicate significant differences at 0.01 level.
8
Preference Ratio
7 6
Slash Orientation
5
0 degree 45 degree 90 degree 135 degree
4 3 2 1 0 Horizontal (DE)
Vertical (DE)
Pictorial Direction of Elongation
Fig. 2. Interaction between slash orientation and pictorial DE.
4. Discussion The major purpose of this study was to investigate the effects of pictorial solidity, size and DE, and orientation and thickness of the red-circle slash on preference ratings for prohibitive symbols. Data analysis showed that solid pictorials were rated more preferable than outline pictorials. This result is consistent with previous research findings (cf. Sanders and McCormick, 1993). A plausible explanation is that outline figures are visually more complicated because they are constructed with more lines, angles and segments than solid figures (Hochberg and Brooks, 1960). Many studies found that complicated figures degraded human performance (Easterby and Zwaga, 1984; Curry et al., 1998; Dewar, 1999). For pictorial size, preference ratings increased as size increased from 50% to 100%. However, there was no significant improvement in ratings from 75% to 100% pictorial size. Generally speaking, larger pictorials are easier to recognize, and hence would receive better preference. Caird et al. (1997) found that small pictorials and pictorials with slashes result in difficulties in comprehension. This finding also agreed with Matlin’s finding (1989) that when the area hidden by a slash did not exceed a certain amount, symbol recognition was not degraded. Based on the results of this study, a pictorial size that is 75% of the inner diameter of the red circle-slash is appropriate, and 100% is not necessary.
Shieh and Huang (in preparation for publication) had similar finding. They investigated the effects of pictorial size and slash thickness on glance legibility of prohibitive symbols. Their results showed that a pictorial size that is 75% of the inner diameter of the red circle-slash is appropriate under degraded environment. For pictorial DE, the preference ratings for vertical and horizontal pictorial were similar. Pictorials with a greater vertical DE (i.e., tall/thin pictorials) did not significantly differ from pictorials with a greater horizontal DE (i.e., short/wide pictorials) in terms of their preference ratings. However, the significant interaction between the pictorial’s DE and slash orientation indicated that preference rating was lower when the pictorial’s DE and slash orientation were the same. Possibly, this result occurred because when the pictorial’s DE and slash orientation were the same, the slash would obscure more features of the pictorial and perhaps make it harder to recognize. By the same token, a diagonal slash (45 or 135 ) would obscure less of the pictorial than would a slash aligned with the pictorial’s greatest DE and thus be rated more preferable, as found in this study. However, the greater preference ratings for diagonal slashes, even where vertical slashes overlaid horizontal pictorials, or horizontal slashes overlaid vertical pictorials, indicated that other factors might also be involved. For instance, subjects might have been familiar with the existing prohibitive symbols with diagonal slashes, which they see in their daily activities and thus rated the diagonal slashes higher. Thickness of the red circle-slash significantly affected preference ratings. Subjects rated the 25% thickness the best, 35% next, and 45% the worst. Preference ratings decreased as thickness of the red circle-slash increased. Besides, the significant interaction between pictorial size and slash thickness indicated that preference rating increase for 75% or 100% pictorial size leveled off when the thickness was below 35%. However, subjects rated the 25% thickness the best, 35% next, and 45% the worst when the pictorial size was 50%. The evidence indicated that the preference for prohibitive symbols of thicker slash and smaller pictorial size might be degraded
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
586
Table 3 Preference rating scores under combinations of pictorial size and slash thickness Pictorial size Variables
Level
Slash thickness
25% 35% A 45%
50%
75%
100%
Mean (s)
Duncan
Mean (s)
Duncan
Mean (s)
Duncan
4.68 (3.66) 3.78 (2.57) 2.74 (2.21)
A
6.77 (2.71) 6.25 (2.87) 5.39 (3.26)
A
6.70 (2.54) 6.65 (2.61) 6.11 (2.75)
A
B C
A B
A B
Note: Different letters in Duncan group indicate significant differences at 0.01 level.
Preference Ratio
8 7 6 5 4 3
Slash Thickness 25% 35%
2
45%
1 0 50%
75%
100%
Pictorial Size
Fig. 3. Interaction between pictorial size and slash thickness.
drastically. Apparently, the greater the thickness, the lesser the available space for the pictorial and the more the pictorial was covered by the slash, and hence resulted in lower ratings, especially when pictorial size was small. Therefore, to acquire better symbol preference, the slash thickness may be thin enough when the pictorial size is small. However, Subjects may overlook the red slash shape in prohibitive symbol when the slash thickness is not thick enough. In order to keep clear legibility of both pictorials and slashes in prohibitive symbols, the slash size should not be too small. Based on this result, the recommendation of ISO 3864 (1984) that the red circleslash should be at least 35% of the area of the sign may be suitable when the pictorial size is not too small. But, it still needs further study. For slash orientation, ratings for the diagonal (45 and 135 ) slashes were similar, and were rated higher than those for the non-diagonal (0 and 90 ) slashes. This may give the designer of prohibitive symbols some flexibility to use a 45 or 135 slash depending on which would obscure the pictorial to a lesser degree. These results also agreed with Hoffmann et al.’s (1982) findings that slashes with top right to bottom left and slashes from top left to bottom right had no significant differences for glance legibility when circleslashes were over pictorials. However, they also found that slashes with top right to bottom left were better than slashes from top left to bottom right for distance legibility. Therefore, more research is needed to replicate these results.
Finally, the features of shapes are various. Beside vertical DE, or horizontal DE, there are square shape, L shape, T shape, O shape, etc. The six pictorials used in present study cannot represent all the other kinds of pictorial. Whether the results of this study are also valid for other types of pictorials requires further study. The results of this study were very encouraging. Variables that significantly affect the preference rating of prohibitive symbols were identified and their relative effects quantified. Useful follow-on efforts include studies that utilize objective recognition performance measures in tasks that are more closely appropriate realworld tasks, and compare that performance with the predicted preference rating based on the results from the present study.
5. Conclusion Based on the results of the present study, the following conclusions are made concerning the design of prohibitive symbols with respect to preference rating: (1) Solid pictorials are better than outline pictorial. (2) Pictorial size should be at least 75% of the diameter of the inside circle of the prohibitive symbol. (3) The pictorial’s greatest DE and the slash orientation should not be the same. In other words, short/ wide pictorials should not be used with horizontal slashes; nor should tall/thin pictorials be used with vertical slashes. (4) Diagonal slashes (45 or 135 ) are preferred, versus, horizontal or vertical slashes. (5) The thickness of the red circle-slash may be as small as 35% as suggested by ISO 3864 (1984).
Acknowledgements This study was supported by a Research Grant from the National Sciences Council of the Republic of China, Grant No. NSC-89-2213-E-161-011.
ARTICLE IN PRESS K. Shieh, S. Huang / Applied Ergonomics 34 (2003) 581–587
References ANSI, 1991. American National Standard for Environmental and Facility Safety Signs: Z535.2. American National Standard Institute, Washington, DC. Ashley, F., Myer, J., Smith, D., 1971. City signs and lights, Boston Redevelopment Authority, January. Biederman, I., 1987. Recognition by components. Psychol. Rev. 4, 115–147. Bullimore, M.A., Fulton, E.J., Howarth, P.A., 1991. Assessment of visual performance. In: Wilson, J.R. and Corlett, E. (Eds.), Evaluation of Human Work. Taylor & Francis, London, pp. 648–681 (Chapter 26). Caird, J.F., Wheat, B., McIntosh, K.R., Dewar, R.E., 1997. The comprehensibility of airline safety card pictorials. In: Proceedings of the Human Factors and Ergonomics Society 41st Annual Meeting, Santa Monica, pp. 801–805. Curry, M.B., McDougall, S.J.P., Bruijn, O. 1998. The effects of the visual metaphor in determining icon efficacy. In: Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting, Santa Monica, CA, pp. 1590–1594. Dewar, R., 1999. Design and evaluation of public information symbols, In: Zwaga, H.J.G. (Ed.), Visual Information for Everyday Use. Taylor & Francis, London, pp. 285–303. Dewar, R.E., 1976. The slash obscures the symbol on prohibitive traffic signs. Hum. Factors 18 (3), 253–258. Dewar, R.E., Swanson, H.A., 1972. Recognition of traffic control signs. Highw. Res. Rec. 414, 16–23. Easterby, R., 1970. The perception of symbols for machine displays. Ergonomics 13, 145–158. Easterby, R., Zwaga, H., 1984. Information Design: The Design and Evaluation of Signs and Printed Material. Wiley, New York, 588pp.
587
Garner, W.R., 1974. The Processing of Information and Structure. Erlbaum, Potomac, MD. Glover, B.L., Magurno, A.B., Murray, L.A., Wogalter, M.S., 1996. Pictorial Negations: preferences for different circle-slash variations. In: Proceedings of the Human Factors and Ergonomics Society 40th Annual Meeting, Santa Monica, pp. 910–914. Gough, P.B., 1965. Grammatical transformation and speed of understanding. J. Verbal Learn. Verbal Behav. 4, 107–111. Hochberg, J., Brooks, V., 1960. The psychophysics of form: reversibleperspective drawings of spatial objects. Am. J. Psychol. 73, 337–354. Hoffmann, E.R., Smith, G., Kalnins, C., Sanders, B., 1982. Legibility of symbolic turn-restriction signs of various formats. Proc. Aust. Road Res. Board 11 (5), 127–134. ISO 3864, 1984. International Standard for Safety Colours and Safety Signs. ISO, Switzerland. Martindale, C., 1991. Cognitive Psychology: A Neural-Network Approach. Cole, CA. Matlin, M.W., 1989. Cognition. Holt, Rinehart and Winston, NY. Murray, L.A., Magurno, A.B., Glover, B.L., Wogalter, M.S., 1998. Prohibitive pictorials: evaluations of different circle-slash negation symbols. Int. J. Ind. Ergonomics 22, 473–482. Ringseis, E.L., Caird, J.K., 1995. The comprehensibility and legibility of twenty pharmaceutical warning pictogram. In: Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting, Santa Monica, pp. 974–978. Sanders, M.S., McCormick, E.J., 1993. Human Factors in Engineering Design. McGraw-Hill, New York, pp. 121–125. Wogalter, M.S., Murray, L.A., Glover, B.L., Shaver, E.F., 2002. Comprehension of different types of prohibitive safety symbols with glance exposure. In: Proceedings of the Human Factors and Ergonomics Society 46th Annual Meeting, Santa Monica, pp. 1753–1757.