16
Applying and testing an approach to design for culturally diverse user groups P. Bourges-Waldegg a, * , S.A.R. Scrivener b a Research and Advanced Studies Centre, Mexico b Coventry University, Coventry, UK Received 16 July 1998; received in revised form 21 December 1999; accepted 29 January 2000 Abstract This paper intends to illustrate how user interface designers can apply the Meaning in Mediated Action (MIMA) approach (P. Bourges-Waldegg, A.R. Scrivener, Meaning; the central issue is cross- cultural HCI design, Interacting with Computers, 9 (3) (1998) 287–310, special issue on “Shared Values and Shared Interfaces”) to design for culturally diverse user groups. After outlining its theoretical foundation, we describe how the MIMA stages—observation, evaluation, analysis and design—were carried out to redesign a WWW system. Finally, we assess the efficacy of this approach by comparing the results of the evaluation of the original and the redesigned interfaces. q 2000 Elsevier Science B.V. All rights reserved. Keywords: Culture; Interface design; Representation; Meaning; Context 1. Introduction With developments in communication and computer technologies, a number of new possibilities have arisen for the interaction between people worldwide. For example, geographically dispersed individuals and groups can now work together through the computer, can communicate in new ways, and can interact in many other transactions, from playing a computer game, to exchanging knowledge, or trading with a range of products and services. However, for the interface designer, the use of shared computer systems, such as GroupWare, Computer-Supported Co-operative Work (CSCW), Compu- ter-Supported Communication (CSC), and the Internet presents a great challenge, parti- cularly in how one deals with cultural factors in design, as geographically dispersed user Interacting with Computers 13 (2000) 111–126 0953-5438/00/$ - see front matter q 2000 Elsevier Science B.V. All rights reserved. PII: S0953-5438(00)00029-1 www.elsevier.com/locate/intcom * Corresponding author. Centro de Investigacion y de Estudios Avanzados, Rincon Colonial de Calacoaya, Atizapan Endomex 53996, Mexico. Tel.: 1 52-5-397-7182. E-mail address: [email protected] (P. Bourges-Waldegg).

Applying and testing an approach to design for culturally diverse user groups

Embed Size (px)

Citation preview

Page 1: Applying and testing an approach to design for culturally diverse user groups

Applying and testing an approach to design forculturally diverse user groups

P. Bourges-Waldegga,*, S.A.R. Scrivenerb

aResearch and Advanced Studies Centre, MexicobCoventry University, Coventry, UK

Received 16 July 1998; received in revised form 21 December 1999; accepted 29 January 2000

Abstract

This paper intends to illustrate how user interface designers can apply the Meaning in Mediated

Action (MIMA) approach (P. Bourges-Waldegg, A.R. Scrivener, Meaning; the central issue is cross-

cultural HCI design, Interacting with Computers, 9 (3) (1998) 287±310, special issue on ªShared

Values and Shared Interfacesº) to design for culturally diverse user groups. After outlining its

theoretical foundation, we describe how the MIMA stagesÐobservation, evaluation, analysis and

designÐwere carried out to redesign a WWW system. Finally, we assess the ef®cacy of this

approach by comparing the results of the evaluation of the original and the redesigned interfaces.

q 2000 Elsevier Science B.V. All rights reserved.

Keywords: Culture; Interface design; Representation; Meaning; Context

1. Introduction

With developments in communication and computer technologies, a number of new

possibilities have arisen for the interaction between people worldwide. For example,

geographically dispersed individuals and groups can now work together through the

computer, can communicate in new ways, and can interact in many other transactions,

from playing a computer game, to exchanging knowledge, or trading with a range of

products and services. However, for the interface designer, the use of shared computer

systems, such as GroupWare, Computer-Supported Co-operative Work (CSCW), Compu-

ter-Supported Communication (CSC), and the Internet presents a great challenge, parti-

cularly in how one deals with cultural factors in design, as geographically dispersed user

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 111

Interacting with Computers 13 (2000) 111±126

0953-5438/00/$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.

PII: S0953-5438(00)00029-1

www.elsevier.com/locate/intcom

* Corresponding author. Centro de Investigacion y de Estudios Avanzados, Rincon Colonial de Calacoaya,

Atizapan Endomex 53996, Mexico. Tel.: 1 52-5-397-7182.

E-mail address: [email protected] (P. Bourges-Waldegg).

Page 2: Applying and testing an approach to design for culturally diverse user groups

groups will often be culturally diverse. Traditional approaches, such as internationalisa-

tion and localisationÐhere called culturalisation [1]Ðcannot be used effectively in the

case of systems shared by culturally diverse users because they are based on recognising

the differences that exist between cultures in order to produce speci®c versions adapted to

the needs of a given ªtarget cultureº. Instead, an approach is needed, capable of dealing

directly with culturally heterogeneous user groups, and capable of integrating cultural

diversity, rather than diversifying the user groups into target cultures, because the exis-

tence of several ªtarget culture adaptationsº can complicate user±user interaction instead

of facilitating it. Moreover, since the purposes of a shared-system include those of inter-

action and communication, it can be assumed that the culturally diverse user is prepared to

interact with other users who speak different languages, or with interfaces which utilise

other languages.

According to Bourges-Waldegg and Scrivener [2], culturally determined usability

problems have a common origin in understanding the intended meaning of the representa-

tions used in the system (including those involved in the user's interaction with the task,

the environment, the tool and other users1). They (ibid.) argue that the cultural differences

affecting usability and design are mainly representational, and that a culturally determined

usability problem can be characterised as the user's dif®culty in understanding that repre-

sentation R means M in context C (adapted from Ref. [3]). In order to understand that

representation R means M, knowledge of the context in which that meaning is rooted is

needed. That is mainly because the meaning of a representation is determined by its

context of use, i.e. where and how the representation is used, what are the surrounding

or associated representations and meanings, etc. Interface elements affected by culture,

such as colour, words, numbers and sound [4], appear to be problematic largely because

they are representations and their meanings can be understood differently by culturally

diverse people when they do not share the knowledge of the context in which they are

rooted.

Hence, culturally heterogeneous user groups are particularly dif®cult to deal with, as

their members may not share the knowledge of the contexts in which the intended mean-

ings of representations are rooted, and this can produce dif®culties in understanding,

during system interaction. However, it is important to recognise that culturally diverse

users may also share the knowledge of some of these contexts independently of the culture

they pertain to. In other words, a person does not necessarily need to belong to a speci®c

culture to share a context. For example, the meanings of the Italian words ªpizzaº,

ªmozzarellaº and ªpepperoniº are cross-cultural because they are based on a shared

context determined by the worldwide advertisements of pizza brands. Shared contexts

occur because neither people nor cultures are isolated entities ignorant or blind to the

`different other'. Although a culture can be de®ned as a speci®c group of people sharing a

distinctive set of values, symbols, rituals, heroes, etc., cultures are, to different degrees, the

result of their interactions with other cultures [5].

Thus, from a design point of view, we argue that the problem for user interface

designers primarily is one of how to communicate the functionality of the system to the

members of a culturally heterogeneous user group when representations can be culturally

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126112

1 User-user, user-tool, user-task and user-environment interaction based on the HCI model proposed by [10].

Page 3: Applying and testing an approach to design for culturally diverse user groups

relative and therefore misunderstood. To tackle this problem the designer needs to pose

two basic questions: (1) Will the users understand that representation (R) means (M) in the

context of the interface (C)? (2) Will a culturally heterogeneous user group share the

context (C) in which that meaning is rooted? ªRepresentationº is de®ned as any aspect

of the system conveying or intended to convey meaning, ªmeaningº is what the repre-

sentation conveys, and ªcontextº refers to how and where a representation is used, and to

the representations surrounding it.

In this way, if for example, the designer is working on a tourist information WebSite and

needs to communicate the existence of a link to a restaurants database, then, she/he needs

to:

1. Ask whether the users will understand that the representation chosen to communicate

this function, say a fork-and-knife icon (R) means restaurant (M) in the context of the

webpage interface (C).

2. Determine if the culturally diverse members of the user group share the knowledge of

the context (C)Ðsay, international airport signalingÐin which the fork-and-knife icon

(R) is based.

The ®rst point will serve to establish whether a representation is cross-culturally under-

standable in the interface context. The second point will help to determine if the culturally

diverse users share the context from which the icon and its meaning were taken and

therefore could understand other representations and meanings rooted in the same context.

If the answer in each case is yes, then the culturally diverse users will probably understand

the metaphorÐa link to a restaurant databaseÐas well as other representations rooted in

that context, despite the fact that some may eat using chopsticks.

This interpretation of the interface design problem reveals where the designer of shared-

systems should concentrate, in order to tackle culturally determined usability issues:

1. on evaluating whether the culturally diverse users understand that representation (R)

means (M) in context (C);

2. on determining whether the culturally diverse users share the context (C) in which the

meaning of a representation is rooted;

3. on designing or redesigning representations from identi®ed shared contexts.

We called the approach proposed above `Meaning in Mediated Action', MIMA [2]. The

MIMA approach consists of four cyclic steps: observation, evaluation, analysis, and

design. In the following pages we will describe these stages and how they were applied

to redesign the interfaces of a WWW system consisting of a browser (Netscape English

version 2.0) and a WebSite (ªthe Nemetonº). We selected Netscape's browser for evalua-

tion because although it is widely used and shared we expected culturally determined

usability problems to occur. This expectation was based partly on the fact that Netscape

was originally designed to suit a culturally speci®c user group, and on the results of other

cross-cultural evaluations of interactive systems [6]. We chose ªthe Nemetonº WebSite

(comprising a series of pages about a pop-music band called ªThe Shamenº) because we

considered that although it dealt with a topic of possible cross-cultural interest, it was

designed in a very culturally speci®c way.

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 113

Page 4: Applying and testing an approach to design for culturally diverse user groups

2. MIMA observation

The aim of MIMA observation is to understand how representations mediate the actions

of the users and to identify any breakdowns in this mediation. Hence, this stage consisted

of observing a sample of the target culturally diverse group in interaction with the WWW

system (see Appendix A) in order to identify these breakdowns. According to Ref. [7] a

breakdown entails that something is wrong with the system and therefore the user has to

break it down to elucidate the problem. In the particular case of MIMA we focused on

problems related to the understanding of meaning. For example, we identi®ed dif®culties

in understanding the intended meaning of the representations involved in the WebSite's

Visitor's Book (the English word ªreloadº, the computer speci®c jargon, etc.), which

produced breakdowns in the user±tool and user±task interactions, and consequently in

the user±user interaction. Also in the WebSite, almost all the tags used as links produced

problems. These included ªFeedbackº, ªDiaryº, ªInfoBaseº, ªThe Endº and the ªForumº.

The representations used in the browser also produced breakdowns on the understanding

of both the tool and the task. For example, the functions involving the words ªHandbookº

and ªDirectoryº were misunderstood, even though all of the users spoke English. The

word ªCoolº was a problem too. The following example taken from one observation

session, shows how the meaning of the word ªcoolº is not clear to the user.

J: (¼) for example ªWhat's Coolº does that mean like the temperature, or? Maybe

the pop group, I mean the jargons of the band.

Although he recognises the word and understands some of its meanings, the context in

which the meaning of ªcoolº in Netscape is rooted is unfamiliar to him.

The majority of the breakdowns detected during the MIMA observation occurred in the

user±tool interaction. Therefore, MIMA evaluation focused on the tools' representations

and helped to establish which of the problems described above were general usability

issues involving representations and which were culturally determined usability problems.

3. MIMA evaluation

MIMA evaluation focuses on asking if the users understand that representation (R)

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126114

Table 1

Interviewer record format

BROWSER:

Interviewee: D

Culture: Spanish

Representation Participant's de®nition

Home Starting point

Reload Load again

What's new? Name of an anti-virus program

File Documents, programs

What's Cool? ªCool is Coldº, The button's purpose may be to show ªWhat's wrong

with the program, if it has failedº

Page 5: Applying and testing an approach to design for culturally diverse user groups

means (M) in the context (C) of the interface. Hence, this stage consisted mainly of

interviewing a different sample of culturally heterogeneous users, to ask them about the

meaning of each of the problematic representations detected during the observation stage.

The interviewer recorded in a structured format the participants' de®nitions (see Table 1).

To carry out the interviews, a horizontal prototypeÐinvolving only the main function-

ality of the system [8]Ðwas produced. We did not use the actual system for MIMA

evaluation because we wanted to prohibit possible learning of the representations'

intended meaning gained from using the system. The participant's responses were

audio-recorded in order to backup and con®rm the data from the interviews.

4. MIMA analysis

The aim of MIMA analysis is to assess the extent to which the users understand each

representation and to determine whether the culturally diverse users share the context (C)

in which each representation is rooted in order to inform design or redesign decision

making. This stage was carried out in three ways:

1. By assessing the user's understanding of intended meaning (Table 2); the intended

meaning of each representation is compared with the user's de®nition and the designer

assesses if the user's understanding can produce any misconceptions or problems. If the

user's de®nition is reasonably close to the intended meaning of the representation, then

the evaluator ticks in the ªAssessment of IMº box. If the evaluator believes that the

user's de®nition can be a source for potential misunderstandings, or simply is too

different from the intended meaning, then she/he puts a cross in the ªAssessment of

IMº box.

2. By determining the cultural speci®city of each representation. The speci®city analysis

is done by comparing all of the participants' assessment-records (see Table 3). Those

representations that were widely understood can be seen as pertaining to a shared

context from which other representations can be taken for design purposes. We can

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 115

Table 2

Example of a participant's assessment of intended meaning

Interviewee: D

Culture: Spanish

Representation Intended

meaning

Participant's

de®nition

Assessment

of IM

Home Point of departure Starting point U

Reload Load again Load again U

What's new? What is novel? (what WebSites) Name of an anti

virus program

£

File Organised related material or items.

Metaphor of a box or folder for

organising documents.

Documents, programs U

What's Cool? What is good? (what WebSites,

according to Netscape)

ªCool is coldº, The button's purpose

may be to show ªWhat's wrong with

the program, if it has failedº

£

Page 6: Applying and testing an approach to design for culturally diverse user groups

determine whether a context is shared from this analysis because we set out from the

thesis that in order to understand a representation, knowledge of its context is needed.

Hence, if the users understood a particular representation it means that they know the

context in which it is rooted, and therefore they ªshare itº.

3. By comparing the participants' de®nitions in order to ®nd common use of representa-

tions (see Table 4).

5. MIMA design

Some of the representations that presented problems were selected to illustrate the

MIMA redesign process. It is important to note that design decisions were taken purely

on the basis of culturally determined usability problems, therefore the redesigned versions

may present other kinds of usability problems, but this is not an issue here. If MIMA was

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126116

Table 3

Example of the speci®city of representations

Representations Participants Total Speci®city (%)

A B C D E F G H I J K

Home £ £ £ 3 27.2

Reload £ £ 2 18.1

What's new? £ £ £ £ £ 5 45.4

File £ £ 2 18.1

Cool £ £ £ £ £ £ £ £ 8 72.7

Table 4

Example of the comparison between participants' interpretations

Representation Intended meaning Participants' de®nitions

What's Cool? What is good?

(what WebSites,

according to Netscape)

ªColdº, ªoddº, ªup to dateº, ªnewº, ªWhat's wrong

with the program, if it has failedº

Directory Listing ªDirectorate boardº, ªwho runs the Netscape

companyº, ªtree-style structure sequenceº,

ªdictionaryº, ªa manual providing directions on how

to use the systemº, ªdrive c:/and drive a:/º (as in PCs)

History Record

(or visited WebSites).

This representation was generally interpreted as ªthe

history of the Internet, of Netscape or of the

WebPagesº i.e., how it was born, how it evolved, who

create it, etc. Also as the subject of History e.g., past

events of a country

White pages A metaphor of a Phone Book

(the function includes a directory

to ®nd the names and e-mail addresses

of Internal users).

ªBlank pageº ªNew pageº, ªBlank Documentº,

ªForm or of®cial reportº

Page 7: Applying and testing an approach to design for culturally diverse user groups

integrated within an overall design processÐsuch as the Star Model [9]Ðthese other

problems would be considered, and all aspects of usability would be examined when

making design decisions.

MIMA analysis helped in the redesigned stage mainly in two ways: (1) to determine

which representations needed to be redesigned; and (2) to bring to light shared contexts in

which new representations could be based.

One of the representations that needed to be replaced according to the MIMA analysis

was ªWhat's Cool?º The speci®city analysis shows that the majority of the participants did

not understand it. Redesign was done then by analysing those representations that were

widely shared in order to ®nd other words associated with them (pertaining to the same

context) that could convey the intended meaning. For instance, one of the representations

more widely shared was ªLocationº (usually de®ned as place or site). Since the actual

function of the ªWhat's Cool?º button is to link the user with a page containing recom-

mended sites or locations in the Web, we decided to use the word ªSiteº in the new

representation ªBest Sitesº (Fig. 1). The word ªBestº was chosen simply because it is a

common word in English and the users shared this linguistic context. Using the same

rationale, ªWhat's New?º was substituted by ªNew Sitesº (Fig. 1) in order to give some

consistency to the buttons.

Another example is that of ªHandbookº which was replaced with ªNet Helpº (Fig. 1).

This decision was made taking into account the MIMA analysis information where the e-

mail icon's question mark (Fig. 2) was generally interpreted by the participants as ªhelpº,

which is exactly the purpose of a handbook. Also, the word ªhelpº seems to be more

common in the vocabulary of non-native English speakers than ªhandbookº, as the parti-

cipants used exactly that word to de®ne the question mark. Therefore the new representa-

tion's meaning is rooted in a shared context±possibly international airport signaling±from

which other associated representations can be taken to communicate other system func-

tions.

The case of ªWhite Pagesº is very similar. The majority of the participants misunder-

stood this representation as a function to open a new page or a blank document. However,

its function is to link the user with a search engine to ®nd people addresses. Since the word

ªSearchº in the ªNet Searchº button was shared by all of the participants we decided to

replace ªWhite Pagesº with ªPeople Searchº (Fig. 3). The word ªPeopleº was again

chosen because it is a common word in English, and the users shared this linguistic

context.

The representation ªNetscape Newsº was replaced by ªDiscussion Groupsº (Fig. 4).

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 117

Fig. 1. Browser's directory buttons (redesigned).

Fig. 2. The original e-mail icon.

Page 8: Applying and testing an approach to design for culturally diverse user groups

According to MIMA analysis, ªnewsº (ªnews groupsº) was not culturally shared.

However, the existence of such a concept in the context of the Internet was not alien

for the heterogeneous participants. When analysing the representation ªthe Forumº in the

WebSite, the existence of such a concept in the participants' knowledge became clear

(some of the participants misinterpreted it as a metaphor for discussion groups). Hence, the

representation was modi®ed to communicate the function in that way. Even if the concept

is unfamiliar for some users, the function may be easier to understand and learn with these

words.

6. Testing MIMA

In order to assess the ef®cacy of MIMA, the redesigned representations were subjected

to a MIMA re-evaluation,2 which was carried out a few minutes after the MIMA evalua-

tion of the original representations with the same culturally diverse eleven participants.

This was possible because at the time of the study (MIMA evaluation and MIMA re-

evaluation), the representations were already redesigned according to a previous pilot

study carried out with a different sample. A prototype (see Appendix B) of the same

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126118

Fig. 3. Index menu (redesigned).

Fig. 4. Widow menu (redesigned).

2 Following redesign, a subsequent MIMA observation should be also considered, as it can provide a more

complete understanding of how these redesigned representations mediate the users actions, e.g. insights into how

the users exploit the detection of a shared-context in deriving meaning of representations embedded in the same

context.

Page 9: Applying and testing an approach to design for culturally diverse user groups

form as that used in the MIMA evaluation of the original representations was used to

ensure that the new representations were tested under conditions comparable to those in

which the original versions were tested.

The same interview record format was used to assess the understandability of the

representations' intended meanings. The participants' responses were audio-recorded.

MIMA analysis was then compared to the MIMA analysis of the original representations

in order to measure any improvements in the understanding of intended meaning. In

addition, a questionnaire designed to record the participants' opinion on the clarity of

the original and redesigned interfaces was also carried out.

To analyse the results of this comparison we will deal separately with the browser and

the WebSite as the latter can be seen as providing more room for improvement.

6.1. Browser

An overall improvement in the understanding of intended meaning can be observed in

Fig. 5. Table 5 indicates the `speci®city' of the original representations and Table 6 the

`speci®city' of their new versions.

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 119

Fig. 5. Number of problems in understanding intended meaning.

Table 5

Speci®city of each of the original representations

Representations Participants Total Speci®city (%)

A B C D E F G H I J K

1 ªBookmarksº £ £ £ £ 4 36.3

2 ªAdd Bookmarkº £ £ £ £ 4 36.3

3 ªWhat's new?º £ £ £ £ £ 5 45.4

4 ªWhat's Cool?º £ £ £ £ £ £ £ £ 8 72.7

5 ªHandbookº £ £ £ £ £ £ 6 54.5

6 ªDirectoryº £ £ £ £ £ £ £ £ 8 72.2

7 ªBrowserº £ £ £ £ £ £ £ £ £ £ 10 90.9

8 ªWhite Pagesº £ £ £ £ £ £ £ £ 8 72.7

9 ªKeyº icon £ £ £ £ £ £ £ £ 8 72.2

10 ªE-mailº icon £ £ £ £ £ 5 45.4

11 ªNetscape Newsº £ £ £ £ £ £ £ £ £ 9 81.8

Page 10: Applying and testing an approach to design for culturally diverse user groups

As it can be observed, there is a considerable decrease in both the overall number of

problems and the number of problems with each representation. The word ªNetº was not

understood on some occasions, however, this did not have an effect on the overall under-

standing of the ªNet Helpº button. The word ªSitesº was not understood only by partici-

pants F and G.

The results of the questionnaire were also positive. The great majority of the

participants found the redesigned representations easier to understand, i.e. eight

participants found the redesigned versions more understandable, two participants

found the original interface easier, and one thought that there was no difference

between them.

6.2. WebSite

As in the case of the browser, there was an overall reduction in the number of problems

in understanding intended meaning (see Fig. 6). Problems in understanding the intended

meaning of the redesigned representations were minimal; only three participants failed to

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126120

Table 6

Speci®city of each of the new representations

Representations Participants Total Speci®city (%)

A B C D E F G H I J K

1 ªFavouritesº £ £ 2 18.1

2 ªAdd to Listº £ 1 9.0

3 ªNew Sitesº £ 1 9.0

4 ªBest Sitesº £ 1 9.0

5 ªNet Helpº 0 0

6 ªIndexº 0 0

7 ªNavigatorº £ £ £ £ £ 5 45.5

8 ªPeople Searchº 0 0

9 ªPadlockº icon £ £ 2 18.1

10 ªE-mailº icon 0

11 ªDiscussion Groupsº 0 0

Fig. 6. Number of problems in understanding intended meaning.

Page 11: Applying and testing an approach to design for culturally diverse user groups

understand two of the representations (see Table 8). Respectively, Tables 7 and 8 show the

speci®city of the original representations and of their new versions. Comparison of these

two tables reveals a decrease in the speci®city of each representation.

The questionnaire also shows the improvement of the redesigned version, as the

majority of the participants found it the easiest to understand, i.e. only one participant

found both interfaces to be equally understandable the rest (10) found the redesigned

version more understandable.

7. Discussion

As the same participants undertook both evaluations it might be argued that the results

are explained by learning. However, the second MIMA evaluation was conducted imme-

diately after the ®rst evaluation, therefore there would have been no opportunity for

participants to gain experience of Netscape and the Nemeton WebSite. In addition, the

order in which the representations were tested was not the same in MIMA evaluations 1

and 2. Nevertheless, it can be argued that participants had experience with the ideas behind

the complete interface before they started with the second session. We believe that this is

not the case because the actual functions of the representations were never discussed with

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 121

Table 7

Speci®city of each of the original representations

Representations Participants Total Speci®city (%)

A B C D E F G H I J K

1 ªDiaryº £ £ £ £ £ £ £ £ £ 9 81.8

2 ªInfoBaseº £ £ £ £ £ £ 6 54.5

3 ªAxis Mutatisº £ £ £ £ £ £ £ £ £ £ £ 11 100

4 ªForumº £ £ £ £ £ £ £ £ £ £ £ 11 100

5 ªThe endº £ £ £ £ £ £ £ £ £ £ £ 11 100

6 ªNemeton Radioº £ £ £ £ £ £ £ £ £ £ 10 90.9

Table 8

Speci®city of each of the new representations

Representations Participants Total Speci®city (%)

A B C D E F G H I J K

1 ªConcertsº 0 0

2 ªInformationº 0 0

3 ªNew CDº 0 0

4 ªProjectsº £ £ 2 18.1

5 ªNightclubsº £ 1 9.0

6 ªListenº 0 0

Page 12: Applying and testing an approach to design for culturally diverse user groups

the users and the original and the re-designed representations were very different in their

forms and meanings. For instance, the original representation ªWhite Pagesº and its re-

designed version ªPeople Searchº: the former was interpreted as a blank page or document

by the majority of the participants so it seems dif®cult that the users could have guessed the

meaning of ªPeople Searchº from that interpretation. Other example is ªThe Endº and

ªNightclubsº; the users misunderstood the former as the exit to the site, so it is unlikely

that they could have learned the existence of a link to ªThe End Nightclubº within

the WebSite. Hence, we do not believe that learning was a signi®cant factor in the

results. Nevertheless, we do recommend in future practice the use of different

samples for MIMA evaluation and re-evaluation to avoid the possibility of learning

effects.

It can also be argued that the observed improvements were in¯uenced by the possibility

of participants being more relaxed during the second MIMA evaluation and therefore

more responsive to the evaluator's questions. Since the representations tested at the

beginning of MIMA evaluation 1 were generally understood (i.e. many generated few

problems) we do not believe that this represents a signi®cant factor. Moreover, we

would argue that although participants might have been more relaxed during the

second evaluation they were also likely to be more tired and therefore more likely

to record additional failures to understand the intended meaning of the redesigned

representations.

Finally, the results of the questionnaire measuring the participants' opinions

about the understandability of the original and redesigned interfaces may also be

disputed, as the users could have assumed or guessed that the second interface was

supposed to be better. However, the questionnaire was only used to support the

results of the comparison between MIMA evaluation 1 and 2 that shows that the

redesigned representations were better understood by the culturally heterogeneous

participants.

8. Conclusions

In order to illustrate how the MIMA approach works in practice, we described how each

of its stages was carried out to redesign a WWW system. MIMA observations led to the

detection of problems with the way representations mediated the actions of the users. The

majority of these problems emerged in the user-tool interaction. Hence, MIMA evaluation

focused on these problems by interviewing a sample of culturally heterogeneous users in

order to record their understanding of meaning. The participants' de®nitions were then

assessed and analysed during MIMA analysis. The representations that needed to be

redesigned were identi®ed and speci®c as well as shared contexts, then used to aid the

MIMA redesign process, were detected.

From the practical experience described in this paper, we can also say that using the

MIMA approach, an improvement on the WWW system's usability was achieved. Tested

by a culturally heterogeneous user group the MIMA redesigned representations were more

easily identi®ed and understood. In addition, the majority of users perceived these repre-

sentations as clearer.

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126122

Page 13: Applying and testing an approach to design for culturally diverse user groups

Interculturally shared-systems are becoming ever more popular around the world and

yet there are very few ways of dealing with the culturally determined usability problems

that can occur when using them. MIMA is an attempt to help designers resolve these kinds

of problems, based on including rather than excluding potential users on the basis of

cultural differences. As it was illustrated, it provides a general framework capable of

uncovering the particular culturally determined usability problems affecting speci®c

design cases, and helps designers to understand how their representations mediate the

users' actions. In this way, MIMA also relieves the designer of the task of consulting and

assessing culturalisation guidelines and does not require specialised expertise, other than

having a clear understanding of the concepts here described, i.e. context, representations,

and meaning.

Other more general advantages of this approach are that it can be used along with a

general HCI approach, to ensure that culturally determined usability problems are

addressed within the more general framework of HCI, and that it is likely to be cost

effective in comparison with culturalisation as, for example, it does not require extensive

study of ªtarget culturesº, or the production of several culturally speci®c versions of an

interface [2].

Appendix A. Prototype

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 123

Page 14: Applying and testing an approach to design for culturally diverse user groups

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126124

Ad d Bo o k m a r k

Ne t sca p e Ho m eWh a t ’s Ne w ?Wh a t ’s Co o l ?

Ne t sca p e Gal l e r i aI n t e r n e t Di r e ct o r yI n t e r n e t Se a r chI n t e r n e t Wh i t e Pa g e sAbo u t t h e I n t e r n e t

Ne t sca p e M a i lNe t sca p e Ne w sAdd r e ss Bo o k

Boo k m a r k sHis t o r y

Gen e r a l Pr e f e r e n ce sM a i l a n d Ne w s Pr e f e r e n ce sNe t w o r k Pr e f e r e n ce sSecu r i t y Pr e f e r e n ce s

Sho w To o l b a rSho w Lo ca t i o nSho w Di r e ct o r y Bu t t o n s

Au t o Lo a d I m a g e s

Docu m e n t Enco d i n g

Sav e Op t i o n s

Ba ckFo r w a r dHo m eSto p Lo a d i n g

Und o

CutPa st eCle a r

Sel e ct Al l

Fi n d . . .Fi n d Ag a i n

Ne w M a i l M e ssa g eM a i l Do cu m e n t . . .Ope n Lo ca t i o n . . .Ope n Fi l e . . .

Clo seSav e As . . .Up l o a d Fi l e . . .

Pa g e Se t u p . . .Pr i n t . . .

Qui t

Fi l e Ed i t

Rel o a dRel o a d Fr a m eLoa d Ima g e s

Docu m e n t So u r ceDocu m e n t I n fo

Vi e w

Op t i o n sBo o k m a r k sGo

Di r e ct o r y Wi n d o w

Page 15: Applying and testing an approach to design for culturally diverse user groups

Appendix B. Prototype (redesigned)

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126 125

Page 16: Applying and testing an approach to design for culturally diverse user groups

References

[1] P. Bourges-Waldegg, A.R. Scrivener, Designing Interfaces for Culturally-diverse Users, Proceedings of the

Sixth Australian Conference in CHI, IEEE Computer Society Press, New Zealand, 1996 (pp. 316±317).

[2] P. Bourges-Waldegg, A.R. Scrivener, Meaning; the central issue is cross-cultural HCI design, Interacting

with Computers 9 (3) (1998) 287±310 (special issue on ªShared Values and Shared Interfacesº).

[3] J.R. Searle, The Construction of Social Reality, Penguin, London, 1995.

[4] E.M. Del Galdo, in: J. Nielsen (Ed.), Internationalisation and Translation: Some Guidelines for the Design

of Human±Computer Interfaces, Designing User Interfaces for International Use, Elsevier, Amsterdam,

1990, pp. 1±10.

[5] R. Robertson, Globalization: Social Theory and Global Culture, Sage, London, 1992.

[6] P. Baird, Hypertext-towards the single intellectual market, in: J. Nielsen (Ed.), Designing User Interfaces

for International Use, Elsevier, Amsterdam, 1990, pp. 111±122.

[7] P.C. Wright, A.F. Monk, Evaluation for design, in: A. Sutcliffe, L. Macaulay (Eds.), People and Computers,

V, Cambridge University Press, Cambridge, 1989, pp. 345±358.

[8] J. Preece, Human±Computer Interaction, Addison-Wesley, Wokingham, UK, 1994.

[9] D. Hix, H.R. Hartson, Developing User Interfaces: Ensuring Usability through Product and Process, Wiley,

New York, 1993.

[10] B. Shackel, Usability-context, framework, de®nition, design and evaluation, in: B. Shackel, S. Richardson

(Eds.), Human Factors for Informatics Usability, Cambridge University Press, Cambridge, 1991.

P. Bourges-Waldegg, S.A.R. Scrivener / Interacting with Computers 13 (2000) 111±126126