ARTICLE IN PRESS

0003-6870/$ - se

doi:10.1016/j.ap

�CorrespondE-mail addr

Applied Ergonomics 39 (2008) 209–217

www.elsevier.com/locate/apergo

Exploratory research on reading cognition and escape-routeplanning using building evacuation plan diagrams

Chieh-Hsin Tanga,�, Ching-Yuan Lina, Yu-Min Hsub

aDepartment of Architecture, National Taiwan University of Science and Technology, Taipei, TaiwanbDepartment of Construction Engineering, Chung Hua University, HsinChu, Taiwan

Received 2 January 2006; accepted 1 May 2007

Abstract

The purpose of evacuation plan diagrams is for readers to comprehend and then plan an evacuation route. However, comprehending

such diagrams involves complex issues that have yet to be addressed by research. This study aims to investigate how Taiwanese people

interpret evacuation plan diagrams in their buildings. Issues of interest include the amount of time that it takes for a member of the

general public to read a diagram and the time that they spend planning their escape route. Correlated and influencing factors are

analyzed. The floor plan of an actual department store was used as the diagram for cognitive testing. A method of stimulated

measurement was conducted over the Internet. The results of the experiment showed that the time it takes to plan an escape route is

about 1.1 to 2 times longer than its reading time. This indicates that there is a significant time difference between diagram interpretation

and stimulated planning. It was found that the longer it takes to read a diagram, the longer it takes to plan an escape route. In addition,

to understand the difference between interpretations by the general public versus those with an architectural background, an analysis

showed that the general public takes two to three times longer than architectural professionals to read a diagram and plan an escape

route. Consequently, improvements in reading diagrams could help in the planning of a more efficient escape route. Furthermore,

through our analysis, we found that the design of diagram symbols must satisfy conventional use and also that diagrams must avoid the

use of metaphorical and abstract symbols. Diagrams that follow our guidelines will generally result in more effective and efficient

conveyance of the intended message, thereby assisting in an emergency.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Evacuation plan diagram; Reading cognition; Escaping disaster

1. Introduction

While the display of evacuation plan diagrams in publicplaces is required by law in Taiwan, it is not clear howpeople effectively utilize these diagrams in actual practice.Wogalter et al. (2002) and Shieh and Huang (2003)demonstrated that readers’ understanding of advisorysymbols can be increased by optimizing the layout anddesign of the graphics. With respect to deriving meaningfrom diagrams, Murray et al. (1998) believed that differentcircle-slash negation symbols can produce different levelsof cognitive preference. He investigated whether four typesof circle-slash (a slash over the symbol, a slash under the

e front matter r 2007 Elsevier Ltd. All rights reserved.

ergo.2007.05.001

ing author. Fax: +8862 2737 672.

ess: tang.ch@msa.hinet.net (C.-H. Tang).

symbol, a partial slash, or a translucent slash) differ interms of how people perceive their effectiveness, asmeasured via preference rankings. They found that theover and under slashes were preferable to translucent orpartial slashes. Thus, the form of expression of a symbolicgraphic could influence the results of cognition. Piamonteet al. (2001) investigated symbols across different culturesand found that cultural background influences one’sunderstanding of symbols. Stramler (1993) proposed threekinds of graphic symbols: (a) those that stand for some-thing else; (b) those that communicate a use for an object/structure; and (c) those that communicate what should orshould not be done at a given time or location. Graphicalsymbols, in turn, are usually icons, pictograms, or pictorialsymbols (Bocker, 1993). They use very little space and arenon-text dependent, but they can convey considerable

ARTICLE IN PRESSC.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217210

information (Maguire, 1985). Studies by Easterby (1970)and Sanders and McCormick (1993) revealed that solidfigures were clearly superior to outline figures, but theseinvestigators did not examine prohibitive symbols. Increas-ing the size of pictorials improves legibility, althoughbeyond a certain size legibility attains asymptotic levels andmay even deteriorate (Bullimore et al., 1991). Suchgraphical displays deliver information or meaning throughthe pictorial symbols and evacuation plan diagrams behavein a similar way; such pictograms require certain learningprocesses to be understood.

Carpmen et al. (1984) found that as the number of signsin a hospital hallway increased, wayfinding performancedecreased. Seidel (1983) found that 76% of people who haddifficulty wayfinding in a large metropolitan airport hadtrouble understanding the signs, and 30% of the sample feltthat there were too many signs. Weisman (1987) found thatonly 18% of nursing home residents mentioned the use ofsigns as a strategy for wayfinding. The remaining 82%mentioned using other architectural features as cues fororientation.

Best (1970) identified a positive relationship between thenumber of choice points (hallway intersections) withinbuildings and wayfinding difficulty. Weisman (1981)defined a number of environmental variables that peopleuse to aid in orientation during wayfinding, with fourclasses of variables as: (1) signs, which provide directionalinformation within a setting; (2) perceptual access, whichprovides a view to landmarks within or exterior to abuilding; (3) architectural differentiation, which is the easewith which different regions or landmarks within abuilding can be recognized; and (4) plan configuration,which is the configuration of a building’s floor plan.Beaumont et al. (1984) interviewed building occupants andfound that the layout of floor plans was equal inimportance to other architectural features, such as theavailability of signs, in reported ease of wayfinding. O’Neill(1991a) found that incremental increases in floor plancomplexity reduced the accuracy of the cognitive map andwayfinding performance. O’Neill (1991b) also found thatfloor plan complexity reduced wayfinding performance,despite the presence of directional signage. Nichols et al.(1992) reported that the primary cause of wayfindingdifficulty in transportation centers is the complexity ofpossible paths.

In emergency situations that require locating emergencyexits in an unfamiliar building, the first reaction might beto find a floor plan and figure out the building’s layout andflow of movement. The present study uses evacuation plandiagrams as the target of investigation to understand adiagram’s explanatory power, or how it is able to conveymeaning. We will look at how long it takes for individualsof different backgrounds to read the evacuation plandiagram (the time it takes for them to locate the exit) andhow long it takes for them to plan an emergency exit route(the time it takes them to find their current location and theexit). This will allow us to understand the factors that

influence a diagram’s explanatory power, and whether thediagram can provide sufficient and accurate escape routeinformation.Because evacuation plan diagrams are based on profes-

sional architectural floor plans, another key area of thisresearch was to determine the difference in cognitive resultsbetween individuals of various professional backgrounds.The accuracy of the information in the cognitive map caninfluence wayfinding performance (O’Neill, 1991a, b). Thecognitive map can store ‘route’ and ‘survey’ representations(Tolman, 1948). Route representations contain knowledgeabout individual places, and the way in which they areconnected through experience; the ‘travel-ability’ betweenplaces (Kuipers, 1983). Thus, a basic element for successfulwayfinding is knowledge of the connections between places,since this information is necessary for selection ofsuccessful routes from one’s initial position to one’s final,intended destination, even if one does not know how faraway or in what direction that destination lies (O’Neill,1991a). For the reasons above, this research aims todetermine the effect of professional background oncognitive reading of diagrams. Hence, our specific objec-tives are as follows:

(1)

Using an evacuation plan diagram for cognitive testing,examine the amount of time required to read a diagramand the amount of time required to plan an escaperoute.

(2)

Determine factors, such as gender, age, and profes-sional background, that affect cognitive reading ofdiagrams.

(3)

Through results obtained by addressing the first twoobjectives, establish basic design guidelines for evacua-tion plan diagrams so that they can better assist thegeneral public in an emergency.

2. Methods

2.1. Evacuation plan diagram

The evacuation plan diagram of the fourth floor of adepartment store with eight floors above ground and fivefloors below ground was selected for this study. Theinterior measurements of the floor were approximately80m long by 50m wide, with a floor area of approximately4000m2 Balconies were not included in the area studied inthis research (this is because the design of the escape routeis through the safety stairways, and balconies are notwithin the scope of the emergency evacuation considered).Fig. 1 exhibits the evacuation plan as it appeared in thecomputer model. The areas shown in white representpassages that are available to the participants to moveabout; the areas shown in pale gray represent areas thatcannot be passed through (orange color was used duringexperiment); the four rectangles represent the location ofthe emergency staircases (red color was used during

ARTICLE IN PRESS

Fig. 1. Evacuation plan used in the research.

C.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217 211

experiment); and the red dot represents the current locationof the participant. The participant was only allowed tomove within the white areas (passages), and was notpermitted to cross areas represented by any other color.

2.2. Procedures

A wayfinding task is typically composed of a number ofsmaller problem-solving tasks that culminate in finding thefinal destination. Finding one’s final destination usuallydoes not occur in one step, especially in large and complexsettings. Rather, wayfinding consists of starting from aknown point and attempting to reach an intermediate sub-goal. At this intermediate sub-goal, the user reorients himor herself and decides which direction to take to the nextsub-goal, until the task is complete (O’Neill, 1999).

In order to be able to obtain a random sample ofparticipants with a broad range of backgrounds, onlinecomputer modelling was used to virtually simulate anescape from a building, and diagram interpretation andresponse time were obtained through this method. Com-puter-simulated escape from an emergency situation, usingvirtual reality techniques, was employed by Shih et al.(2000) to compare evacuation time between simulated andtraditional methods for evacuation planning in Japan andTaiwan. The results confirmed that virtual reality techni-ques can be used to acceptably simulate evacuations frombuildings. Therefore, this research uses the Internet tocollect participants’ background information, interpreta-tion ability, and response time. Time recording began assoon as the participants started to read the evacuation plandiagram, and stopped when each participant finishedplanning his or her route. During this time, a participantcould use the mouse to control the red dot within thepassages shown in white to carry out the planning ofevacuation routes. Not until the red dot reached one of theemergency staircases (one of the red rectangles) was theescape considered complete. Time was shown on thedisplay throughout the simulation. Each participant wasallowed only one try, so as not to affect experimentalresults by the participant becoming familiar with thediagram. Furthermore, the participants were not allowedto watch others as they were tested.

2.3. Participants

To broaden the background of the participants sampled,emails were randomly sent with a link to the Web site fortesting diagram reading and route planning. Those whoreceived such an email were asked to not only participate inthe study but also to forward the email to theiracquaintances. Through this process, participants of awide range of background were tested. The study includeda total of 113 participants. Excluding those with incom-plete data (lack of background information), 92 valid datarecords were collected. The ratio of men to women was 2:1,mainly because in Taiwan, more men than women use theInternet. The largest age group was that of 26–35 year olds,who comprised 33.6% of the sample. The overall averageage was 31. In terms of educational level, the largest groupwas that of college-educated people, making up 82% of theentire group, because the focus of this study was todetermine factors that influence diagram cognition, soincluding participants with a professional architecturalbackground was of extreme importance. That is why, inaddition to using random sampling, people with aprofessional architectural background were selected toparticipate, resulting in 62% of all subjects having majoredin architecture or related fields.

2.4. Data analysis

The following statistical methods were used to carry outdata analysis: (1) For diagram reading time and route-planning time, the mean value was used for initialunderstanding of the effect of different demographicvariables on test results. In addition, by adopting the ideaof a trend line, an analysis was made of the relationshipbetween the time required for reading the diagram, route-planning time, and the total time required. R2 (coefficientof determination) was used in the analysis to show how aprofessional background would influence the time requiredfor reading the diagram. (2) To determine what factorsaffect graphical cognition, independent sample t-test usinggender and professional background was conducted. Age,however, required the use of an ANOVA test. (3) Ananalysis was made of the association between the timerequired for reading the instruction map and planning theescape route, to explore the relationship between readingand taking actions. (4) Then, multiple regression analysiswas used to establish a regression function with fourindependent variables of: educational background, profes-sional background, time required for reading the diagram,and route-planning time. Systemized coefficients were thenused to determine how the four factors influenced therequired total time. All statistical analyses were conductedusing the Statistical Package for the Social Sciences (SPSS).Besides, communication theory was adopted in analyzingthe diagram’s communication abilities (the coding anddecoding of the diagram) to establish principles fordesigning the diagram.

ARTICLE IN PRESSC.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217212

3. Results and discussion

3.1. Relationship between participant characteristics and

graphical cognition

The experimental results, categorized by the participants’characteristics, are summarized in Table 1.

The results show that men took less time than women tointerpret diagrams (men: women ¼ 13.89:33.65). It tookwomen three times as much time as men to read diagramsand twice as much time to plan a route. This leads to apreliminary conclusion that Taiwanese men have bettergraphical cognitive ability than women.

In terms of age, younger participants took less time thanolder ones to interpret the diagram and plan their escaperoute. In addition, the time difference between age groupsappears to increase with increasing age, although thedifference is not significant. In his Cognitive Theory, Piagetemphasized that cognitive differences do not reside in age,but in education and learning. The current study alsofound, through experimentation, that graphical cognitiondoes not vary significantly with age.

Table 1 shows that, on average, people whose profes-sions are related to architecture are better at interpretingdiagrams than the general public. In reading diagrams,professionals need half the time that the general publicneeds, indicating that education and learning havesignificant effects. In terms of cognition, this possiblyindicates that professionals have more schemata. There-fore, they can quickly identify a similar diagram schemaand assimilate it with the new diagram to facilitate diagraminterpretation. That is, they can exhibit a greater ability totake accurate action through their interpretation ofdiagrams. In addition, from the (dotted) regression lines,we have the following two equations:

y ¼ 0:592xþ 7:2932,

R2 ¼ 0:5296 ðnon�architecture�related fieldÞ, ð1Þ

y ¼ 0:863xþ 3:7025;

R2 ¼ 0:8189 ðarchitecture�related fieldÞ, ð2Þ

Table 1

Mean time, results according to participants

Background categories Time spent

reading

diagram (s)

Time spent

planning

route (s)

Total

amount of

time (s)

Entire group 8.47 11.02 19.49

Men 5.70 8.19 13.89

Women 15.47 18.18 33.65

I. Age 16–25 5.05 12.21 17.26

II. Age 26–35 8.48 9.62 18.10

III. Age 36–45 11.70 11.35 23.05

Architecture-related

field

5.54 8.49 14.03

Non-architecture-related

field

13.17 15.09 28.26

where y is the route-planning time (in seconds), and x is thetime required for reading the diagram (in seconds). A positiveassociation can be found between the time needed to readdiagrams and the time needed to plan one’s escape route.Moreover, from R2, we can conclude that the results fromparticipants with an architectural background are moreconsistent than for the general public (0.818940.5296). Fora more detailed discussion on ways to reduce differences indiagram cognition, see later. Garling et al. (1983) found thatpeople with restricted visual access to their destination withina building took significantly longer times to learn the layoutthan people with visual access. However, when given access toa map of the floor plan, the group with low visual accesslearned as fast as people with visual access. This researchfurther found that those with architectural knowledge are ableto more quickly capture the information on a diagram. Thegeneral public, on the other hand, has more difficultyinterpreting diagrams. By improving the way diagrams arepresented, the variation in time required to interpret suchdiagrams can be significantly reduced (Fig. 2).

3.2. Analysis of factors affecting differences in graphical

cognition

In order to understand differences in people’s graphicalcognition, participants’ backgrounds were analyzed throughindependent sample t-tests, using gender and professionalbackground as classification. Age was analyzed using anANOVA test. The results are presented in Tables 2 and 3.From Table 2 it is clear that there are statistically

significant differences by gender and by professionalbackground for both diagram-reading and route-planningtimes, men and architecture students being faster. Thesignificance of the time required for diagram readinginvolves locating the position through the recognition ofthe signs and signals on the diagram, i.e. the ability tocomprehend the diagram. On the other hand, the timerequired for route planning signifies the time spentplanning the escape route and direction, based on theperson’s comprehension of the diagram. This can beconsidered as the process during which comprehension istransformed into actions. It thus appears that the (faster)method used by those with a special professional back-ground to comprehend the diagram differs from how otherpeople comprehend the diagram. The recognition differ-ence between the two genders does not necessarily meanthat males tend to be better at understanding the diagramthan females because inequality of the number of malerespondents and female respondents (2:1) and differentialrepresentation by professional background can influencethe outcome of the test. Therefore, it is recommended thatan investigation into how gender influences the recognitionof the diagram should be carried out on a larger scale.

3.2.1. Information provided by evacuation plan diagrams

Evacuation plan diagrams are a form of architecturaldrawing. The way that they are constructed is very similar

ARTICLE IN PRESS

Table 2

T-test results of differences in times required

Variables Time categories Levene test

significance

t 95% confidence interval

Lower limit Upper limit

Gender (male vs. female) Diagram-reading time 0.000� �4.840 �15.4537 �4.0922

Route-planning time 0.003� �2.722 �17.6777 �2.3031

Professional background (architecture-related

field vs. non-architecture-related field)

Diagram-reading time 0.000� 3.268 2.8417 12.4251

Route-planning time 0.001� 2.107 0.1832 13.0177

�Significance levelo0.05, indicating that the difference between the values is statistically significant.

y = 0.592x + 7.2932

R2 = 0.5296

y = 0.863x + 3.7025

R2 = 0.8189

0

5

10

15

20

25

30

35

40

45

50

0 10

diagram-reading time (s)

rout

e-pl

anni

ng ti

me

(s)

non-architecture-related field architecture-related field

regression line (non-architecture-related field) regression line (architecture-related field)

20 30 40 50 60

Fig. 2. Comparison of graphical cognition time between groups of different backgrounds.

Table 3

ANOVA results of time differences according to age

Time categories Comparison

groups

Level of

significance

Diagram-reading time I vs. II 0.080

I vs. III 0.005�

II vs. III 0.246

Route-planning time I vs. II 0.172

I vs. III 0.197

II vs. III 0.992

Note: 1. I ¼ age 16-25; II ¼ age 26–35; III ¼ age 36–45 2.�Level of significanceo0.05, indicating that a difference between values

is statistically significant.

C.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217 213

to architectural drawings. They include content withsymbolic meaning. These drawings follow the conventionsshared by architecture and related fields. To professionals

in these fields, the symbols have standardized meanings.For lay people though the symbols lack conventionalmeaning, and therefore are not publicly aware standar-dized symbols. Fiske (1990), in his Theory of Communica-

tion, proposed three main paths for standardizing asymbol: by convention and use; by explicit agreement;and by implicit clues. Standardization by explicit agree-ment is closely related to reading cognition of evacuationplan diagrams. Explicit agreement refers to the clear,agreed-upon relationship between the signifier and thesignified. There is a clear, shared understanding amongusers of these symbols. These kinds of text and graphicsneed to be learned, because cognitive differences originatefrom differences in cognitive structures, and cognitivestructures, in turn, are formed through education andlearning. In terms of interpreting the symbols in anevacuation plan diagram, two issues need to be addressed.One issue addresses how the design of the symbol (the

ARTICLE IN PRESSC.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217214

symbol’s form of encoding) could conform to sharedconventions to convey its meaning effectively. The otherissue is how cognitive interpretation can be achievedthrough education and learning (the process of decodingthe symbol) and, thus, how one can obtain a shared,accurate schema and develop a shared understanding of theinformation in the diagram. The meaning conveyed by thesymbols in a diagram is not derived from the relationshipbetween the reader and the symbol. This information hasbeen embedded into the text through clear, agreed-uponconventions. In other words, it is a replacement relation-ship. Many elements in the evacuation plan diagrams havea similar feature. In the process of finding an exit route, oneencounters several symbols, such as those for doors, exits,and elevators. In case of an emergency, one needs to usethese symbols to derive meaning. For professionals, thesymbols follow clear, agreed-upon conventions and, there-fore, are easily understood.

Text-based symbols function by way of metaphors,which do not rely on clear logic or rules, and therefore areunsuitable for use in evacuation plan diagrams. Metaphorsare especially unsuitable for reading and interpreting incases where one is trying to escape danger. Thus, symbolicelements must be removed from evacuation plan diagramsas the general public lacks schema to interpret evacuationplan diagrams. The inability of the general public to gathersafety information from the diagrams is likely to result indifferent interpretations and more time required forplanning an escape route.

Given that information transmission follows a knownroute, only if the symbols used in evacuation plan chartsconform with conventional ones can they transmit safetyinformation effectively. Currently, the law specifies re-quirements regarding the content of these diagrams (e.g.signs for escape routes, location of exit doors andemergency stairs, location of emergency equipment), butdoes not specify how these elements should be expressed inthe diagrams. Given that the general public does notpossess the required schema and a shared understanding,effective transmission of emergency escape informationbecomes a major issue when designing evacuation plandiagrams.

In addition, related research indicates that total disor-ientation and the sense of being lost can be a frighteningexperience and can lead to severe emotional reactions(Zimring, 1982; Carpmen et al., 1986). Some of thesereactions, if rationally assessed, may appear exaggerated,but they tend to signify more than just temporary spatialuncertainty. They evoke feelings of anxiety and insecurity;they affect self-esteem and judgement of self-competence(Passini, 1999). Wayfinding ability in complex buildingsalso affects security during emergency evacuations. Tosummarize a complex field of study, it can be said thatsettings which operate well under normal circumstances,especially if their floor layout is easily understood and ifthey offer clear alternatives of exiting, also will be safer inemergencies. The observation that people in emergency

situations tend to use routes and exits they know (Canter,1980; Sime, 1985) indicates that the often half-hidden,never-used emergency exits are not the answer to buildingsafety. Under stressful situations, such as in the event of anevacuation, the ability to interpret emergency evacuationplans is further crippled, making the importance of adiagram that conveys its information effectively moreapparent.

3.2.2. Relationship between diagram interpretation and

reflective action

In addition to investigating how people understand themeaning of diagrams, we should further examine the time ittakes for people to interpret diagrams and put what theyhave learned into action in case of an emergency. Asindicated in Table 2, the general public may need moretime to interpret the symbols in diagrams than experts.Passini (1999) selected hospitals as an example for theirwayfinding experiments and found that most patients hadproblems decoding symbols and abbreviations on graphicdisplays. The spatial disposition of messages on signs alsoled to confusion. Signs were not understood because thearrow was too far away from the destination name, whileindependent messages were associated and interpreted asone because of close proximity (Passini et al., 1999). That isattributed to the fact that people could not identify asimilar or related schema in their mind to facilitateinterpretation of the diagrams and immediately plan andexecute an escape route.Participants with an architecture-related background

needed much less time than those with other backgroundsto read the diagram and plan an exit route. Whether anincrease in the complexity of the floor plan would widenthe time gap between those two groups of participants andsignificantly affect understanding of space and directionduring an actual emergency deserves attention. Theinformation provided to the general public in evacuationplan diagrams might not be sufficient, or people mightchoose merely to follow others, because the diagrams aredifficult to interpret. This diminishes the value andeffectiveness of the diagrams.As shown in Table 3, there were age related significant

differences only between age groups I and II. This resultindicates that relevant differences in cognition are relatedless to age than to the level of education and learning, atleast at the young and middle-aged levels in this experi-ment.The relationship between diagrams and reading inter-

pretation can be illustrated as in Fig. 3. Through reading,people communicate with the evacuation plan diagram anduse the derived information as a basis for action in case ofan emergency. This communication process consists ofencoding and decoding steps.

Encoding is the way in which evacuation plan diagramsare constructed. Designers convert emergency exit plansinto pictorial and textual explanations, as expressed in theevacuation plan diagram. We suggest that evacuation plan

ARTICLE IN PRESS

transmission of

meaning

original

meaning

interpreted

interpretation

of original

meaning

emergency

exit diagram

(encoding)

reading

(decoding)

required characteristics:

1. comprehensibility

2. ease to be conventionally understood

3. clarity

4. able to give directions

5. safety concepts

cognitive process

assimilation to present

schema

OK Nor

accommodation

cognitive balance

person object

original meaning

new meaning

interpretedoriginal meaning

not interpreted

Fig. 3. Relationship between diagram and reading.

C.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217 215

diagrams should include the following characteristics: (1)comprehensibility; (2) ease of being conventionally under-stood; (3) clarity; (4) ability to provide directions; and (5)safety concepts. These five features together must conveythe intended meaning. Encoding can be done in the form ofarchitectural floor plans, but the diagrams need to havecommonalities such as ‘‘convention and use’’ and ‘‘explicitagreement’’ in John Fiske’s (1990) communication theory.People rely on a variety of architectural cues (atriums,elevators, variations in floor and wall color, visual access tooutside landmarks, plan configuration, etc.) to find theirway through buildings. These architectural features form asystem of landmarks (O’Neill, 1999) and should thereforebe included in the list of items that need to be consideredwhen encoding diagrams.

Decoding is a cognitive process. Readers must use thesymbolic elements of a diagram together with words tointerpret it. A person must go through the process ofassimilation, accommodation and balance to interpretmeaning. In order to improve the general public’s sharedgraphical schema for decoding, we need to rely ongovernment educational campaigns. O’Neill (1999) alsobelieves that a suitable degree of familiarity with diagramshelps users to understand interior architectural design.Lastly, the intersection of what is encoded and decoded isthe meaning interpreted, and that area must equal theoriginal meaning. Only in that case is an evacuation plandiagram considered to have achieved its task of transmit-ting information accurately.

Thus, there are two recommendations for improving theeffectiveness of diagrams. One recommendation is toimprove the form of expression used in the diagram, bydecreasing the number of technical symbols. The otherrecommendation is to increase the general public’s abilityto interpret diagrams through education and learning.

3.2.3. Relationship between required times

In the experiment reported here diagram interpretationtime is the combination of diagram reading and routeplanning. The relationship between these two processesdeserves further exploration in order to understand therelationship between a diagram’s ability to inform and thepeople’s ability to interpret directions. The results arepresented in Tables 4.For those with a professional background, the time they

spend on reading the diagram and planning the escaperoute show a rather low non-significant level of correlationat 0.229, showing that the existing professional schema hasbeen functioning, enabling clear understanding of thediagram (locating the current position based on thediagram) and recognition of escape measures to be taken(finding a escape route to the exit). For non-professionals,the correlation value of the two time durations is 0.558,significant at 0.006, so that the longer the time required fordiagram reading, the longer the time needed for theplanning of an escape route.To explain the contribution of each variable to the

total time required, a multiple regression analysis wasconducted. The results are shown in Table 5. Allvariables in the model remained below the 0.05 levelof significance, revealing that the results of the analysiswere statistically significant. We know from the Betacoefficient that time required for route-planning (Be-ta ¼ 0.676) influences the total amount of time themost. Level of education was negatively correlated tothe total amount of time; the higher educated a personis, the less time is required. Professional backgroundexhibited a similar effect although because professionalbackground data was not categorized in terms of yearsand the number of participants was limited, furtherinvestigation is needed.

ARTICLE IN PRESS

Table 4

Relationship between diagram-reading and route-planning

Categories of people Time spent reading

diagram (s)

Time spent

planning route (s)

Correlation

coefficient

Level of

significance

General public (diagram reading vs. route planning) 13.17 15.09 0.558 0.006

Professionals (diagram reading vs. route planning) 5.54 8.49 0.229 0.172

Table 5

Influence of each variable on total amount of time

Variable Standardized

coefficient (Beta)

Level of

significance

Educational background �0.401 0.001

Professional background �0.385 0.001

Diagram-reading time 0.532 0.000

Route-planning time 0.676 0.000

C.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217216

4. Conclusions

4.1. Time required for diagram interpretation

In general, planning an exit route takes about 1.1 to 2times the amount of time taken to read an evacuation plandiagram. This suggests a gap between diagram interpreta-tion and action planning. During cognitive processing,assimilation and accommodation are not the same asschema recognition. The time required for assimilation islonger than for schema recognition. The general publictakes twice the amount of time taken by a professional toread a diagram and plan an exit route. This indicates thatdiagrams contain professional content that is not easilyunderstood by lay people.

4.2. Factors affecting time required for diagram

interpretation

When the general public interprets diagrams, the longerthe time it takes to read, the longer it takes to plan a route.However, among professionals, this is not the case. Thereason is that informational diagrams rely on conventionsto convey their message, and architectural professionalspossess more diagram schemata than lay people, allowingthem to quickly and accurately interpret the content ofsuch diagrams. Personal background is the main factorinfluencing diagram-reading and route-planning. Route-planning is the main factor influencing the total amount oftime. This implies that there is considerable difficulty inactually transferring understanding of a diagram intoaction.

4.3. Special characteristics that diagrams should possess

To be able to form a cognitive consensus, emergencyevacuation diagrams must be able to fulfill special

characteristics of conventional use. Moreover, the use ofprofessional symbols should be minimal. However, makingthe signifier and signified clear and creating a directrelationship between the two (so that one can besubstituted for the other) is only achieved throughlearning/training. Furthermore, metaphorical and abstractsymbols must be eliminated to more effectively convey theintended message and assist in an emergency.

References

Beaumont, P., Gray, J., Moore, G., Robinson, B., 1984. Orientation and

wayfinding in the Tauranga Departmental Building: A focused post-

occupancy evaluation. In: Duerk, D., Campbell, D. (Eds.), Environ-

mental Design Research Association Proceedings, vol. 15. San Luis,

CA, Obispo, pp. 77–91.

Best, G.A., 1970. Direction-finding in large buildings. In: Canter, D. (Ed.),

Architectural Psychology. RIBA Publications, London, pp. 72–91.

Bocker, M., 1993. A multiple index approach for the evaluation of

pictograms. Proceedings of the 14th International Symposium on

Human Factors in Telecommunications. R. v. Decker’s Verlag, G.

Schenck GmbH, Heidelberg, pp. 73–85.

Bullimore, M.A., Fulton, E.J., Howarth, P.A., 1991. Assessment of visual

performance. In: Wilson, J.R., Corlett, E. (Eds.), Evaluation of

Human Work. Taylor & Francis, London, pp. 648–681 (Chapter 26).

Canter, D., 1980. Fires and Human Behavior. Wiley, Chichester, UK.

Carpmen, J.R., Grant, M.A., Simons, P.A., 1984. No More Mazes:

Research About Design for Way Finding in Hospitals, Patient and

Visitor Participation Project. University of Michigan Hospitals, Ann

Arbor, MI.

Carpmen, J.R., Grant, M.A., Simons, P.A., 1986. Design that cares.

American Hospital Association, Chicago.

Easterby, R., 1970. The perception of symbols for machine displays.

Ergonomics 13, 145–158.

Fiske, J., 1990. Introduction to Communication Studies. Routledge,

London, New York.

Garling, T., Lindberg, E., Mantyla, T., 1983. Orientation in buildings:

effects of familiarity, visual access, and orientation aids. J. Appl.

Psychol. 68, 177–186.

Kuipers, B.J., 1983. The cognitive map: could it have been any other way?

In: Acredolo, L., Pick, H. (Eds.), Spatial Orientation: Research,

Application, and Theory. Plenum, New York.

Maguire, M.C., 1985. A review of human factors guidelines and

techniques for the design of graphical human–computer interfaces.

Comput. Graphics 9 (3), 221–235.

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. Ergon. 22, 473–482.

Nichols, F., Canete, I., Tuladhar, S., 1992. Designing for Pedestrians: a

CAD network analysis approach. In: Kalay, Y.E. (Ed.), Evaluating

and Predicting Design Performance. Wiley, New York, NY.

O’Neill, M.J., 1991a. Evaluation of a conceptual model of architectural

legibility. Environment and Behavior 23 (3), 259–284.

O’Neill, M.J., 1991b. Effects of signage and floor plan configuration on

wayfinding accuracy. Environment and Behavior 23 (5), 553–574.

ARTICLE IN PRESSC.-H. Tang et al. / Applied Ergonomics 39 (2008) 209–217 217

O’Neill, M.J., 1999. Theory and research in design of ‘You Are Here’

maps. Visual Information for Everyday Use. Taylor & Francis,

London, 225–238.

Passini, R., 1999. Wayfinding: backbone of graphic support systems.

Visual Information for Everyday Use. Taylor & Francis, London, pp.

241–256.

Passini, R., Joanette, Y., Rainville, C., Marchand, N., 1999. Wayfinding

and information processing in dementia. Visual Information for

Everyday Use. Taylor & Francis, London, pp. 267–273.

Piamonte, D.P.T., Abeysekera, J.D.A., Ohlsson, K., 2001. Understanding

small graphical symbols: a cross-cultural study. Int. J. Ind. Ergon. 27,

399–404.

Sanders, M.S., McCormick, E.J., 1993. Human Factors in Engineering

Design. McGraw-Hill, New York, pp. 121–125.

Seidel, A. 1983. Way finding in public spaces: the Dallas-Fort Worth

Airport. In: Armedeo, D., Griffin, J., Potter, J. (Eds), Proceedings of

the 14th Annual Meeting of the Environmental Design Research

Association, Washington, DC.

Shieh, K.K., Huang, S.M., 2003. Factors affecting preference ratings of

prohibitive symbols. Appl. Ergon. 34, 581–587.

Shih, N.-J., Lin, C.-Y., Yang, C.-H., 2000. A virtual-reality-based

feasibility study of evacuation time compared to the traditional

calculation method. Fire Saf. J. 34, 377–391.

Sime, J., 1985. Movement towards the familiar: person and place

affiliation in a fire entrapment setting. Environ. Behav. 17, 692–724.

Stramler Jr., J.H., 1993. The Dictionary for Human Factors/Ergonomics.

CRC Press, London, 413pp.

Tolman, E., 1948. Cognitive maps in rats and man. Psychol. Rev. 55,

189–208.

Weisman, G.D., 1981. Evaluating architectural legibility: wayfinding in

the built environment. Environ. Behav. 13, 189–204.

Weisman, G.D., 1987. Improving way-finding and architectural legibility in

housing for the elderly. In: Regnier, V., Pynoos, J. (Eds.), Housing the

Aged: Design Directives and Policy Considerations. Elsevier, New York.

Wogalter, M.S., Conzola, V.C., Smith-Jackson, 2002. Research-based guide-

lines for warning design and evaluation. Int. J. App. Ergon. 33, 219–230.

Zimring, C., 1982. The built environment as a source of psychological

stress: impacts of buildings and cities on satisfaction and behavior. In:

Evans, G.W. (Ed.), Environmental Stress. Cambridge University

Press, Cambridge, UK.