A Study of Search User Interface Design Based on Hofstede’s Six

Cultural Dimensions

Karen Chessum1 a, Haiming Liu1 b Ingo Frommholz2 c

1School of Computer Science and Technology, University of Bedfordshire, Luton, UK.
2School of Electronics and Computer Science, University of Southampton, UK
School of Engineering, Computing and Mathematical Sciences, University of Wolverhampton, Wolverhampton, UK

[email protected], [email protected], [email protected]

Keywords: Cross-Cultural Information Retrieval, Cross-Cultural Theory, Website Design, Human-Computer

Information Retrieval (HCIR), Hofstede’s Cultural Dimensions, Human-Computer Interaction (HCI).

Abstract: An information seeker’s cultural background could influence their preference for search user interface (UI)
design. To study cultural influences Geert Hofstede’s cultural dimensions have been applied to website
design for a number of years. In this paper, we examine if Hofstede’s six cultural dimension can be applied
to inform the design of search engine user interfaces. The culturally designed search user interfaces have
been evaluated in a study with 148 participants of different cultural backgrounds. The results have been
analysed to determine if Hofstede’s cultural dimensions are appropriate for understanding users’ preferences
on search user interface design. Whilst the key findings from the study suggest Hofstede cross-cultural
dimensions can be used to model users’ preferences on search interface design, further work is still needed
for particular cultural dimensions to reinforce the conclusions.

1 INTRODUCTION feelings and actions which have been learned over

one’s lifetime, once learned have to be unlearned
A user’s cultural orientation could influence their before new patterns can be absorbed. Hofstede,
preferences on user interface (UI) design, as noted by (1991, p 5) also noted, “Culture is learned not
(Reinecke et al., 2010). Contemporary search engines inherited”. Culture is different from human nature on
such as Google, Bing, and Yahoo!, do not one hand and from personality on the other hand.
differentiate between different user types or the Culture as described by Hofstede (1991, p 5) is “the
cultural backgrounds of their users. Research collective programming of the mind which
conducted by Slone (2002, p 1166 ) states, “Both distinguishes the members of one group or category
motivation and experience, elements of goals and of people from another”.
mental models, played equally strong roles in this In our work, we examine cross-cultural aspects in
result. In fact, goals and mental models work in search UI design. The overall aim of our research is
tandem to determine overall searching behaviour”. A to study the potential differences and different
user’s cultural background influences their mental preferences between cultures when it comes to search
model formation. UI. Due to the reported importance of Hofstede’s
One means of defining culture is mental work in international communication, (Wardrobe,
programming. Mental programming can be thought 2005) international management, (Bing, 2004)
of as patterns of thinking, feeling and actions, based international marketing, (Mooij and Hofstede, 2010)
upon what we have learned throughout our lifetime. and use in website design, (Marcus and Gould, 2000:
Hofstede (1991, p 4) writes, “Much of it has been Liu, 2021), we base our considerations on Hofstede’s
acquired in early childhood”. Thinking patterns, model.

a https://orcid.org/0000-0002-2980-8746
b https://orcid.org/0000-0002-0390-3657
c https://orcid.org/0000-0002-5622-5132
 Our contribution is as follows: firstly, we discuss Hall identified the Primary Message Systems
how we have used Hofstede’s six cultural dimensions (PMS). These systems are non-lingual ways in which
to inform our prototype search UI design. Secondly, humans communicate with one another. Hall
in our study we use the prototype UIs we designed identified 10 PMS each relating to a facet of human
based on Hofstede’s dimensions to determine if activity, (Hall, 1990). However, it is Hall’s ‘high-
Hofstede’s cultural dimensions and index scores for context’ and ‘low-context’ work that is most cited
different countries match the user preference of the within a Human-Computer Interaction (HCI)
search UI design. For example, if Hofstede’s cultural perspective.
dimensions Index scores indicate a country’s culture
is towards the high end of the Masculinity dimension, 2.2 Trompenaars & Hampden-Turner
do users from this cultural background actually prefer
a search UI designed with high Masculinity in mind? Trompenaars and Hampden-Turner also looked at
To accomplish our aims, the remainder of the culture from a dimension level and defined seven
paper is structured as follows: to justify our choice of cultural dimensions. They took Parson’s five
applying Hofstede’s dimensions, we briefly review relational orientations, Parsons (1951) as a starting
different cultural models in the next section. point.
Subsequently, we discuss by example how These dimensions are units that can be used to
Hofstede’s dimensions informed the design of our make comparisons and are as follows: Universalism
prototype UIs. Next, we present our study and its vs Particularism, Individualism vs
results to answer the question of whether Hofstede’s Communitarianism, Specific vs Diffuse, Neutral vs
index scores can be used to indicate the search UI Emotional, Achievement vs Ascription, Sequential
preferences of users from different cultural time vs Synchronous time and Internal direction vs
backgrounds. Finally, we offer our conclusion. Outer direction.

2.3 Nisbett
2 CULTURAL MODELS
Nisbett, a social psychologist, examines the
Several cultural models have been critically reviewed differences between Eastern and Western cultures. As
for the suitability of this research, which is to noted by Oshlyansky (2007), Nisbett looks at the
effectively model different cultures and be able to “processes of thought, perception, attention,
inform the design of cross-cultural search UIs. Below organisation of knowledge, understanding” and other
are the details on what they are and why we decide to mental processes. He uses ‘holistic’ and ‘analytic’
continue our investigation with Hofstede’s model. thought patterns or mental processes to distinguish
between Eastern and Western cultures, with the West
2.1 Hall on the analytics side and East/Asian on the holistic
side. Nisbett and Miyamoto, (2005) says, “the
Edward Hall, an anthropologist, was a pioneer in evidence indicates that people in Western cultures
cross-cultural business communication. Hall (1976) focus on salient objects and use rules and
defined culture as using ‘high context’ (HC) and ‘low categorization for purposes of organizing the
context‘(LC). A high context communication, as environment. By contrast, people in East Asian
noted by Smith et al., (2004) is where, “little has to be cultures focus more holistically on relationships and
said or written because most of the information is similarities among objects when organizing the
either in the physical environment or within the environment.”
person, while very little is in the coded, explicit part
of the message”. Liu (2021) notes ‘people from high- 2.4 Hofstede
context cultures prefer face-to-face communication’
and continues by saying high-context cultures ‘look Geert Hofstede, a Dutch anthropologist, carried out
for both less-direct verbal and subtler nonverbal cues in-depth interviews with hundreds of IBM employees
during the communication’. in 53 countries. He identified four cultural
This high-low context for cultures refers to how dimensions and published his research at the end of
information is stored and flows. Whereas in a ‘low the 1970s. The fifth dimension Long-term Time
context’ culture the information contained in the Orientation (LTO) was added in 1991 from work
message is explicit, little is hidden. carried out by Michael Harris Bond, supported by
Hofstede. An additional sixth dimension, Indulgence
versus Restraint (IND) added in 2010 is described in researched extensively in relation to HCI by the
Hofstede et al., (2010) following the analysis of the authors of this paper. We considered the following
World Values Survey data (from the World Values interpretations of the six dimensions (user interface
Survey Organisation) from 93 countries by Michael designs UI 1 to UI 12 4 ) applicable to Human-
Minkov. Hofstede et al., (2010) contains details of all Computer Interaction (HCI) design.
six dimensions. There has been generally less research conducted
regarding Hofstede’s most recent, sixth dimension,
2.5 Why Hofstede? Indulgence v Restraint (IND). Many works conducted
did not include this last dimension. Examples are
Geert Hofstede’s cultural research is probably one of (Marcus and Gould, 2000), (Smith et al., 2004),
the most well-known cultural models and as noted by (Chessum et al., 2014) and (Karreman et al., 2016).
Ghemawat and Reiche, (2011) the most widely used. Hofstede’s additional sixth dimension is included in
Hofstede created six dimensions by which cultures this research and the descriptions and references
can be compared, (Reid, 2015). Hofstede’s work has given below (Sections 3.1 to 3.6).
been used in a number of disciplines, examples of
which are global branding and advertising’, (Mooij 3.1 Power Distance (PD)
and Hofstede, 2010), consumer behaviour, (Milner et
al., 1993), management control systems, (Chatterjee, Power Distance is the amount of unequal power
2014), cross-cultural psychology, (Hofstede, 2011) within a culture that members of that culture are
and cross-cultural HCI research (Smith et al., 2004). prepared to accept or expect. User interface designs
As noted by Oshlyansky (2007) when discussing for UI 1 and UI 2 are given below:
cultural models used in HCI states, “By far the most
popular of these models is Hofstede’s”. Smith et al., 3.1.1 UI 1 High (PD)
(2004) also state, “Hofstede’s (1991) dimensions of
culture that are the most often quoted theories in  Images of Experts, official buildings, official
relation to cross-cultural usability”. Mooij and logos, prominence given to security and
Hofstede (2010) assert “People perform information- restrictions (Marcus and Gould 2000).
seeking tasks faster when using web content created  Structured website design (Burgmann et al.
by designers from their own culture”. Mooij and 2006).
Hofstede (2010) go on to say, culturally adapted  “Older people are both respected and feared”
websites are more usable, and users are more likely to use images of older people for wisdom and
have a more positive outlook toward them. credibility (Hofstede 2011).

3.1.2 UI 2 Low (PD)

3 HOFSTEDE’S CROSS  Status is displayed to leaders rather than the
CULTURAL THEORY population, staff or consumers. Information
hierarchy is shallow (Marcus and Gould, 2000).
Examples of how Hofstede’s work has been used in  Use a looser structure to allow users to explore
HCI, are shown by Marcus and Gould (2000), who your site for themselves (Nahai, 2013), (Marcus
used Hofstede’s five-dimensional model to develop a and Gould, 2000).
set of website design guidelines and Smith et al.,  Use earned evaluations e.g. ratings,
2004) who incorporated Hofstede’s dimensions in testimonials, likes, to promote your goods or
their process model. Research undertaken by services (Nahai, 2013).
Chessum et al., (2014) looked at Hofstede’s five  “Older people are neither respected nor feared”
dimensions to group users for cross-cultural Show images of younger or youthful people
information retrieval. However, since then, Hofstede (Hofstede, 2011).
has added a sixth dimension, called ‘Indulgence v
Restraint’ (IND), to his cultural model. 3.2 Individualism (IDV)
This work examines whether Hofstede’s
dimensions can be used in the design of search UIs. Individualism within a culture is where the individual
The attributes of Hofstede’s six dimensions have been is expected only to take care of themselves and their

4 Examples of UI 1 to UI 12 can be seen at

https://github.com/ifromm/cross-cultural-ui-designs/
immediate family. There is no expectation for them Navigation focused on exploring but also on
to take care of anyone else, as opposed to a control (Marcus and Gould, 2000).
collectivist society, where members take care of  Masculine societies are competitive. Motivated
extended families and other group members. User by achievement, heroism, assertiveness, and
interface designs for UI 3 and UI 4 are given below: materialism (Idler, 2013).
 “Admiration for the strong” (Hofstede 2011).
3.2.1 UI 3 High (IDV)  Bright contrasting colours (Voehringer-Kuhnt,
2002), (Dormann and Chisalita, 2002).
 “High text-to-image ratio”. Avoid cluttered
graphics. Show positive images of goal 3.3.2 UI 6 Low (MAS)
achievement (Gould et al., 2000).
 “Create competitions and challenges to engage  User attention is obtained by the use of poetry,
your customers”. “Give visitors a sense of aesthetics, and appealing to uniting values
personal achievement to motivate actions”. (Marcus and Gould, 2000).
Have content that has ‘novelty’ and ‘difference’  Provide contact information and be prepared for
in order to ‘attract attention’ (Nahai, 2013). feedback and questions. “This group is very
 Have their own personal goals. Follow their cooperative and if they want to give feedback,
likes and dislikes (Sinha, 2014). they don’t hesitate to get in contact with you”.
 “Speaking one's mind is healthy” (Hofstede  “Feminine societies are consensus-oriented”.
2011). With a preference for values, corresponding to
cooperation, modesty, care for the weak, and
3.2.2 UI 4 Low (IDV) quality of life (Idler, 2013).
 “Sympathy for the weak” (Hofstede, 2011).
 “High image-to-text ratio” (Gould et al., 2000).  Pastel colours, low saturation (Voehringer-
 Transparency, give users full disclosure, for Kuhnt, 2002, (Dormann and Chisalita, 2002).
example how their data would be used. “Show
that you respect privacy and security of personal 3.4 Uncertainty Avoidance (UA)
info”. “Engage the community – ‘we’ not ‘me’”
(Nahai, 2013). Uncertainty avoidance relates to the extent a culture
 Emphasis on social and organisational goals. is either comfortable or uncomfortable with
 An individual’s goals are less important (Gould uncertainty or unknown situations. User interface
et al., 2000). designs for UI 7 and UI 8 are given below:
 Members of a collective society, aspire to
achieve their in-groups’ goals (Sinha, 2014). 3.4.1 UI 7 High (UA)
 “Harmony should always be maintained’
(Hofstede, 2011).  Tries to show/predict the results or effects of
actions before the user acts. Navigation
3.3 Masculinity (MAS) structures are designed to help prevent users
from becoming lost. Any ambiguity can be
With this dimension, Hofstede refers to gender roles decreased by the use of “Redundant cues”, e.g.
rather than to physical gender. Masculine roles design, sound visual aids (Burgmann et al.,
consist of assertiveness, toughness and competition. 2006).
Masculine work objectives incorporate “earnings,  Simplicity, with clear metaphors, restricted
recognition, advancements and challenge”, as noted options, and limited volume of data (Marcus
by Marcus (2002). While feminine roles are and Gould, 2000).
traditionally ones with an emphasis on caring for the  “The uncertainty inherent in life is felt as a
home, family/children, people and tenderness are continuous threat that must be fought”. “Need
considered prevalent. User interface designs for UI 5 for clarity and structure” (Hofstede, 2011).
and UI 6 are given below:
3.4.2 UI 8 Low (UA)
3.3.1 UI 5 High (MAS)
 Information is maximised by the use of colour
 User attention obtained by games and coding, typeface, font, and sound. Use multiple
competitions. Work tasks, roles, and skills, links but not redundant cueing. Limited control
quick results obtained for limited actions. over navigation e.g. Links could open content
 in new windows that lead away from the  “Service to others is an important goal”. “Social
original webpage(s). Complexity with spending and consumption” (Hofstede, 2011).
maximum content and options. Acceptance of
exploring and risk (can even be encouraged), 3.6 Indulgence vs Restraint (IND)
with a stigma on “over- protection.” (Marcus
and Gould, 2000). The additional sixth dimension, relates to happiness,
 “The uncertainty inherent in life is accepted and freedom of expression and feeling in control of your
each day is taken as it comes”, “Comfortable own life, (Hofstede, 2011). User interface designs for
with ambiguity and chaos” (Hofstede, 2011). UI 11 and UI 12 are given below:

3.5 Long-term Time Orientation 3.6.1 UI 11 High (IND)

(LTO)
 Use and encourage user-generated content.
Also known as, Long-term Orientation versus Short- “Make interactions fun”. “Reflect loose gender
term Normative Orientation (LTO). This dimension roles by using a range of models” (Nahai,
was identified later by Hofstede and Bond (1984), 2013).
where Bond had a questionnaire re-designed, with a  People from an Indulgent culture have a
Chinese culture bias. This he called the Chinese value tendency to put an emphasis on individual
survey, (CVS). LTO is a Confucian philosophy, happiness and wellbeing. Their leisure time is
where members value long-term gain over short-term more significant and people experience more
gain. User interface designs for UI 9 and UI 10 are freedom and “personal control” (MacLachlan,
given below: 2013).
 Maintaining order in the nation is not given a
3.5.1 UI 9 High (LTO) high priority. A perception of personal life
control. Freedom of speech is seen as important
 Offer ways for the user to save browsing (Hofstede, 2011).
history, e.g., wish lists. Together with means of
sharing on social media. Persons with long-term 3.6.2 UI 12 Low (IND)
orientation decisions are comprehensive and
grounded “for the future” (Idler, 2013).  Frugal, show how they can save money.
 Patience shown in attaining results and reaching “Emphasise how you serve the community”.
goals. “Relationships as a source of information “Strict, cultured gender roles” (Nahai, 2013).
and credibility” (Marcus and Gould, 2000).  People from a restrained culture do not display
 “Perseverance in achieving results” (Makkonen, positive emotions as easily, with freedom,
2012). happiness and leisure time not assigned the same
 “Thrift and perseverance are important goals”. significance (MacLachlan, 2013).
“Large savings quota, funds available for  Higher number of police officers per 100,000
Investment” (Hofstede, 2011). population. A perception of helplessness: what
happens to me is not my own doing. Freedom of
3.5.2 UI 10 Low (LTO) speech is not a primary concern (Hofstede,
2011).
 Users require quick results that are consistent
with known values and traditions. Persons with
a short-term orientation would appear “to live 4 SEARCH INTERFACE DESIGN
more in the past and in the present than in the
future” (Idler, 2013). As discussed above, the characteristics of Hofstede’s
 Persons from a very short-term oriented culture six cultural dimensions have been used to create
e.g. Spain have a tendency “to live in the twelve prototype UIs. These consist of two UIs
moment” (Nahai, 2013). created for each dimension, with one interface set for
 A wish for instant results and achieving goals. the lower end and one for the higher end of each
“Rules as a source of information and dimension. (e.g., “high masculinity”, “low
credibility” (Marcus and Gould, 2000). masculinity”).
 “Focus on achieving quick results” (Makkonen,
2012).
 social media
organisations.
Admiration for the A general masculine
strong” (Hofstede, 2011) ‘look and feel’.
Bright contrasting Bold colours such as red,
colours. blue, dark blue and black
(Voehringer-Kuhnt, have been used for
2002), ‘Quick’ search textual
(Dormann and Chisalita, links. With a contrasting
2002) white for ‘About Us’,
‘Sign up’ and ‘Log In’.
With black textual links
Figure 1: Prototype High Masculinity (MAS) UI 5. for ‘Privacy’, ‘Terms’
and ‘Settings’.
Here we offer two examples of how Hofstede’s
Masculinity dimension, both, high MAS shown in
Figure 1 and low MAS shown in Figure 2, have been
applied to the search user UI design5. The prototype
UI design was constructed using the design features
described in 3.3 above. How the design features are
implemented are shown below in tables 1 and 2
respectively.

Table 1: High Masculinity.

High (MAS) HCI HCI Design Feature
Design Feature Implementation UI 5
User attention obtained This has been achieved Figure 2: Prototype Low Masculinity (MAS) UI 6.
by games and by, showing text links for
competitions. ‘Competitions’ and Table 2: Low Masculinity.
(Marcus and Gould, ‘Latest Games’
Low (MAS) HCI Design HCI Design Feature
2000)
Feature Implementation UI 6
Work tasks, roles, and ‘Quick’ search textual
User attention is obtained A general ‘softer’
skills, quick results links are provided,
by the use of poetry, appearance with more
obtained for limited offering quick results for
aesthetics, and appealing and larger images and
actions. popular searches.
to uniting values. icons.
(Marcus and Gould,
(Marcus and Gould.
2000)
2000)
Navigation focused on ‘Quick’ searches and
Provide contact Two ‘About Us’ links
exploring but also search links to ‘Web’,
information and be and a ‘Feedback’ link.
control. ‘Images’, ‘Video’ and
prepared for feedback
(Marcus and Gould, ‘News’
and questions.
2000)
(Idler, 2013)
Masculine societies are A non-cluttered interface
“Feminine societies are A ‘Community Matters’
competitive. Motivated with textual links to
consensus-oriented”. link and imagery
by achievement, heroism, News, Weather and
With a preference for showing multicultural
assertiveness, and Latest movies. Also,
values, corresponding to inclusion. ‘Family’ link
materialism. textual links for
cooperation, modesty, and imagery.
(Idler, 2013) ‘Competitions’ and
care of the weak, and
‘Latest Games’.
quality of life.
Only graphical image
(Idler, 2013)
icons are for ‘YouTube’,
“Sympathy for the weak” A general softer, less
‘Twitter’ and
(Hofstede, 2011) masculine aesthetic ‘look
‘Facebook’, allowing
and feel’
faster access to these

5 Examples of UI 1 to UI 12 can be seen at

https://github.com/ifromm/cross-cultural-ui-designs/
 Pastel colours, low The use of pastel colours participant’s nationality and the index scores for
saturation. for background, Hofstede’s six dimensions.
(Voehringer-Kuhnt, ‘Weather’ icon, ‘News’
2002), (Dormann and icon, Logo, and textual
Chisalita, 2002) system links, such as
5.3 Hypothesis
‘About Us’, ‘Sign up’,
‘Log in’, ‘Feedback’, Our six hypotheses have been developed from both
‘Privacy’, ‘Terms’ and the low and high aspects of Hofstede dimensions and
‘Settings’ are detailed below:
H1: Higher PD Countries will show a preference for
This paper examines the data obtained from the UI 1 and Lower PD Countries for UI 2.
research of the twelve prototype search UIs, designed
using Hofstede’s six cultural dimensions. H2: Higher IDV Countries will show a preference for
UI 3 and Lower IDV Countries for UI 4.
H3: Higher MAS Countries will show a preference
5 USER EVALUATION for UI 5 and Lower MAS Countries for UI 6.
H4: Higher UA Countries will show a preference for
Participants of this survey have been taken mostly UI 7 and Lower UA Countries for UI 8.
from the staff, the current student base and the alumni H5: Higher LTO Countries will show a preference for
of the University of Bedfordshire, England. However, UI 9 and Lower LTO Countries for UI 10.
this survey was also promoted internationally via H6: Higher IND Countries will show a preference for
social media by a number of University staff. The
UI 11 and Lower IND Countries UI 12.
survey was completely anonymous. The data for this
study was collected via an online survey. Our survey
was constructed using the following two sections. 5.4 Data Analysis

5.1 Survey Section One The data was collected via our survey as described in
5.2 where the user was asked to select a preference
This section consists of potentially seven questions for one of the UI pairs. The participant’s nationality
depending upon the user’s responses. This collected is collected in section one of the survey as described
general background information about the participant in 5.1 above. Hofstede’s country index scores have
and consisted of closed questions relating to gender, been applied to participants who identify with the
age, occupation, and culture most identified with. In matching nationality. These index scores are applied
addition, there were several questions relating to for each nationality for all six dimensions.
languages spoken and place of residence. Once the data had been collected, a quantitative
data analysis tool has been to analyse the data.
5.2 Survey Section Two
In this section of the survey, the participant/user is 6 EVALUATION RESULTS
exposed to the twelve prototype UIs, these consisting
of two for each of the six dimensions and are We had 148 participants who completed our survey,
sequentially numbered User Interface (UI) 1 to User made up of 101 participants who at present are
Interface (UI) 12. residing in the UK and 47 who are residing overseas.
The users were asked to pick their preferences We had 97 male and 51 female participants.
from the twelve UIs designed using the above design The 148 participants are from 33 countries.
features (Sections 3.1 to 3.6). The search UI’S were Unfortunately, many of the countries only had 1 to 2
paired for each of the six dimensions, with one UI respondents and as such we have not included their
being the low-end design and the other UI being the results. Likewise, several countries not having an
high-end design for all six dimensions, i.e. two UIs Index score calculated by Hofstede, have also been
per dimension making twelve in total. The user could excluded. The countries and cultures results analysed
only pick one for each pair, as the responses are are as follows, the number of participants given in
mutually exclusive. This paper is based upon the brackets. U.K. (Great Britain) (51) Germany (21),
findings from this section together with the Poland (3), Pakistan (10), Nigeria (5), Bangladesh
(3), Ethiopia (3), China (6), Nepal (5), Sri Lanka (3) According to Hofstede’s Index score, the UK,
and India (7), with 117 participants in total. Germany, Poland, Nigeria, Ethiopia and China are
considered high masculine countries – we would
6.1 Analysis of Preferences for Each expect a preference for search UI 5 for these
User’s Culture countries. This is confirmed for the UK, Germany and
Ethiopia, whereas preferences are mixed for China
6.1.1 Hypothesis 1 Power Distance and India and not confirmed for Nigeria.
Pakistan, Bangladesh and India are considered
Both Germany and the UK are considered to be low close to the central point for a MAS country. Indeed,
PD countries (both Hofstede’s country’s Index scores no clear preference for either of the two search UIs
35), which means a culture that supports the concept could be determined by our data as would be
that inequalities in their society be kept to a expected.
minimum. Therefore, the expectation for H1 (Section Nepal and Sri Lanka are countries with a low
5.3) would be that such countries prefer search UI 2. MAS index score. While the data does not allow for
This was not found in our data, with only 12% (UK) determining a clear preference, there is a slight
resp. 14% (Germany) prefer UI 2. tendency towards the (expected) search UI 6.
Pakistan would be considered close to the central
point for a PD country with an Index score of 55. 6.1.4 Hypothesis 4 Uncertainty Avoidance
Here, H1 (Section 5.3) has been partially supported
with 80% of users showing a preference for UI 1 and Hypothesis H4 in 5.3, says that higher UA Countries
20% for UI 2. will show a preference for UI 7 while lower UA
From the countries with a high PD according to Countries will show a preference for UI 8.
their Index score, our data confirmed the expected Poland, Pakistan, Germany and Bangladesh are
preference for UI 1 for Poland, Nigeria, Ethiopia, considered countries with high uncertainty
China, Nepal, Sri Lanka and India. avoidance. Participants from these countries indeed
Against their expected preference, participants exhibited a preference for UI 7, as expected.
from Bangladesh seem to not prefer UI 1. However, Sri Lanka, Nigeria and Ethiopia are countries
we have to note that we only had a few participants considered to be in the middle range when it comes to
from these countries. the UA score. We would not expect any clear
preference for any of the two UIs. However, this has
6.1.2 Hypothesis 2 Individualism not been confirmed in the case of Nigeria whose
participants exhibit a preference for UI 8. However,
UK, Germany, and Poland are considered Ethiopia with an index score of 55, did show a
countries with a high IDV score, which, according to preference for UI 7.
our hypothesis H2 in 5.3, means UI 3 would be The UK, China, Nepal and India are countries
preferred. with a low UA score. Except in the case of Nepal, our
While this is supported in the case of Poland, we data disagrees with the expected preference for UI 8.
observe this is not the case for the UK and Germany.
India would be considered close to the central 6.1.5 Hypothesis 5 Long-term Time
point for an IDV country with an index score of 48. Orientation
H2 has not been supported with 86% of users showing
a preference for UI 4 and 14% for UI 3. Long-term time orientation cultures value virtuous
Pakistan, Nigeria, Bangladesh, Ethiopia, China, behaviour, perseverance and patience for achieving
Nepal, and Sri Lanka are low IDV countries goals and results. Hypothesis H5 suggests that higher
according to Hofstede’s Index scores, accordingly we LTO Countries will show a preference for UI 9 and
would expect a preference for UI 4. This is confirmed, lower LTO countries will exhibit a preference for UI
except for Bangladesh and Ethiopia. design 10.
Germany and China are considered countries with
6.1.3 Hypothesis 3 Masculinity a high long-term orientation, according to Hofstede’s
Index score. However, our data do not confirm the
Hypothesis H3 (Section 5.3), suggests that higher expected preference for search UI 9.
MAS Countries will show a preference for UI 5 and Countries in the middle range when it comes to
lower MAS Countries will show a preference for UI LTO are the UK, Pakistan, Bangladesh, India and Sri
6. Lanka. We can argue the expectation of no clear
preference for either UIs is confirmed (though partially confirmed and Masculinity having 5
Pakistan shows a slight tendency for UI 9). confirmations with 4 partially confirmed. It indicates
Countries with a low LTO score are Poland and there is a stronger link between specific cultural
Nigeria, both exhibiting the expected preference for dimensions and search UI design. This aspect
UI 10. requires further research, possibly using the 3 most
There is no LTO index score reported by Hofstede popular dimensions and more participants.
for Ethiopia and Nepal.

6.1.6 Hypothesis 6 Indulgence ACKNOWLEDGMENTS

For the Indulgence (IND) dimension, our hypothesis We would like to thank all the participants for
H6 states that higher IND countries will show a contributing to this user study.
preference for UI 11 while lower IND countries will
show a preference for search UI 12.
Nigeria and the UK are both considered high
indulgence countries. Based on our data, we regard REFERENCES
H6 to be confirmed due to the preference for UI 11.
Germany, Bangladesh, India and China have a Bing, J.W. (2004) Hofstede's consequences: The impact of
his work on consulting and business practices, An
low IND score according to Hofstede’s Index.
Executive Commentary by John W. Bing, Academy of
Preferences for these countries are rather mixed, so Management Executive, February 2004, Vol. 18, No. 1.
H6 is not confirmed in these cases. However, it has Burgmann, I., Kitchen, P., and Williams, R. (2006) Does
been confirmed for Poland, also considered a low Culture Matter on the Web? Marketing Intelligence &
IND county. There is no IND index score reported by Planning (24:1), 62-73.
Hofstede for Ethiopia, Nepal and Sri Lanka. Chatterjee, D. (2014) Management Control Systems and
Hofstede’s Cultural Dimensions: An Empirical Study
of Innovators and Low Innovators. Global Business
Review. 15(3) 565–582.
7 DISCUSSION AND Chessum, K., Liu, H., and Frommholz, I. (2014) Applying
CONCLUSION Cross-cultural theory to understand users’ preferences
on interactive information retrieval platform design. In
Proceedings euroHCIR.
Overall, we have found 33 out of the possible 61
Dormann, C., and Chisalita, C. (2002) Cultural Values in
results that support Hofstede’s Index scores. A Web Site Design, in Proceedings of the 11th European
further 12 results were difficult to fully categorise but Conference on Cognitive Ergonomics, Catania, Italy,
partially support the hypothesis. 16 results do not September 8-11.
support the respective hypothesis. Ghemawat, P., and Reiche, S. (2011) National Cultural
Firstly, Hofstede’s Index scores do not address Differences and Multinational Business, Globalization
search UIs but had a different focus originally. One Note Series.
of our contributions is exactly to verify to what degree Gould, E., Zakaria, N., and Yusof, S. (2000) Applying
these scores can be applied to search UIs. It seems, Culture to Website Design: A Comparison of
from our data, that Hofstede’s Index scores show the Malaysian and US Websites. Paper Presented at the
potential to inform search UIs, but they also show Proceedings of IEEE Professional Communication
further research needs to be carried out to shed some Society International Professional Communication
light on the reasons why we get at times inconclusive Conference and Proceedings of the 18th Annual ACM
results and how we can better inform culturally aware International Conference on Computer Documentation.
search UI design. We consider our study as an Technology and Teamwork 162-171.
Hall, E.T. (1976) Beyond Culture. Doubleday, Garden City,
important contribution to triggering this discussion.
New York.
Secondly, we had a limited number of participants
Hall, E.T., (1990) The Silent Language. New York: Anchor
from some countries, although our study attracted Books.
many participants in particular from the UK (51), Hofstede, G. (1991) Cultures and Organization: Software
Germany (21) and Pakistan (10). However, looking of the Mind, Paperback Edition published 1994,
at the data, it would seem the most supported HarperCollinsBusiness, An imprint of
dimension across the countries reported in this study HarperCollinsPublishers, 77 - 85 Fulham Palace Road,
is Long-term Time Orientation, with 7 from 9 Hammersmith, London W6 8JB.
countries (2 countries did not have Hofstede Index Hofstede, G. (2011) Dimensionalizing Cultures: The
scores for this dimension) being confirmed. With Hofstede Model in Context. Online Readings in
Power Distance having 6 confirmations and 2
 Psychology and Culture, 2(1). Nahai, N. (2013) How to Sell Online to Short vs Long-Term
https://doi.org/10.9707/2307-0919.1014 Cultures. Psychology Today.
Hofstede, G., and Bond, M. (1984) The Need for synergy https://www.psychologytoday.com/gb/blog/webs-
among Cross-cultural Studies. Journal of Cross- influence/201308/how-sell-online-short-vs-long-term-
Cultural Psychology 15. (4), 417-433 December 1984. cultures
Hofstede, G., and Minkov, M. (2013) Value Survey Module Nahai, N. (2013), How to Sell Online to Different Cultures:
2013 Manual. https://geerthofstede.com/wp- Power Distance. Psychology Today.
content/uploads/2016/07/Manual-VSM-2013.pdf https://www.psychologytoday.com/gb/blog/webs-
Hofstede, G., Hofstede, G.J., and Minkov, M. (2010) influence/201306/how-sell-online-different-cultures-
Cultures and Organizations: Software of the mind: power-distance
intercultural cooperation and its importance for Nisbett, R.E., and Miyamoto, Y. (2005) The influence of
survival. The McGraw Hill Companies. USA. culture: holistic versus analytic perception. Department
Idler, S. (2013) How to Design for a Cross-Cultural User of Psychology, University of Michigan, Ann Arbor, M1
Experience (part 2/2) 23 April 2013. 48109-1109 Trends in Cognitive sciences Vol 9 No 10.
https://usabilla.com/blog/designing-for-a-cross- October 2005.
cultural-user-experience-part2/ Oshlyansky, L. (2007) Cultural Models in HCI: Hofstede,
Ingwersen, P., and Järvelin, K. (2005). The turn: Affordance and Technology Acceptance. PhD Thesis,
integration of information seeking and retrieval in Swansea University.
context. Secaucus, NJ, USA: Springer-Verlag New Parsons, T. (1951) The Social System. New York: The Free
York, Inc. Press, 1951.
Karreman, J., Romeo, P., Li, Q. (2016) Cross-Cultural HCI Reid, L. (2015) The Importance of Hofstede’s Dimensions
and UX Design: A Comparison of Chinese and Western of Culture. https://sites.psu.edu/global/2015/04/25/the-
User Interfaces working paper 2016. importance-of-hofstedes-dimensions-of-culture/
Liu, F. (2021) Modify Your Design for Global Audiences: Reinecke, K., Schenkel, S. and Bernstein, A. (2010).
Crosscultural UX Design. Nielson Norman Group. Modeling a User’s Culture. The Handbook of Research
https://www.nngroup.com/articles/crosscultural- in Culturally-Aware Information Technology:
design/ Perspectives and Models.
MacLachlan, M. (2013) Indulgence vs. Restraint – the 6th Sinha, J.B.P. (2014) Psycho-Social Analysis of the Indian
Dimension. https://www.communicaid.com/cross- Mindset, 2, Springer India 2014 27. DOI: 10.1007/978-
cultural-training/blog/indulgence-vs-restraint-6th- 81-322-1804-3.
dimension/ Slone, D. J. (2002). The influence of mental models and
Makkonen, E. (2012) Cultural differences and localization goals on search patterns during web interaction. Journal
in user interfaces. Thesis for Bachelor of Engineering of the American Society for Information Science and
(B.E.), Media technology. Central Ostrobothnia Technology, 53(13), pp. 1152–1169.
University of Applied Sciences. Smith, A., French, T., Dunckley, L., Minocha, S., and
Marcus, A., and Gould, E. (2000) Crosscurrents Cultural Chang, Y. (2004). A Process Model for developing
Dimensions and Global Web User-Interface Design. Usable cross-cultural websites, Interacting with
Interactions, (7) 4, 32-46. DOI= Computers: Special Edition – Global human-computer
http://dl.acm.org/citation.cfm?doid=345190.345238. systems cultural determinants of usability. 16, 1 (Feb
Milner, L., Fodness, D., and Speece, M. (1993) Hofstede's 2004), 63-91.
Research on Cross-Cultural Work-Related Values: Trompenaars, F., and Hampden-Turner, C. (1998) Riding
Implications For Consumer Behavior. European the Waves of Culture. Understanding Cultural
Advances in Consumer Research, 1, 70-76. Diversity in Business. (2nd.Ed.) Nicholas Brealey
http://acrwebsite.org/volumes/11610/volumes/e01/E- Publishing Limited, 36 John Street, London. WC1N
01 2AT.
Mooij, M., and Hofstede, G. (2010) The Hofstede model Voehringer-Kuhnt, T. (2002) Kulturelle Einflüsse auf die
Applications to global branding and advertising Gestaltung von Mensch-Maschine Systemen, Munich:
strategy and research. International journal of GRIN Verlag.
advertising 29(1), 85-110. Wardrobe, W. (2005). Beyond Hofstede: Cultural
Nahai, N. (2013) How to Sell Online to Individualist vs applications for communication with Latin American,
Collectivist Cultures. Psychology Today. Association for Business Communication Annual
https://www.psychologytoday.com/gb/blog/webs- Convention.
influence/201307/how-sell-online-individualist-vs-
collectivist-cultures
Nahai, N. (2013) How to Sell Online to Indulgent vs
Restrained Cultures. Psychology Today.
https://www.psychologytoday.com/us/blog/webs-
influence/201308/how-sell-online-indulgent-vs-
restrained-cultures