Investigating the Application of Hofstede's Model of ... - DigiTool Menu [PDF]

Investigating the Application of Hofstede’s Model of Cultural Dimensions to Arabic Web Interfaces

Nouf Khashman School of Information Studies, Faculty of Education McGill University

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy

©2014 Nouf Khashman

Abstract Culture is argued to be one of the attributes affecting the usefulness and usability of websites. Localizing a website by incorporating culturally appropriate design features helps it to become both more attractive and more functional for its users. The bulk of research in this domain has utilized the influential cultural model of Geert Hofstede. Based on data collected from more than 50 countries and regions, Hofstede proposed a model consisting of five dimensions of national culture by which individual countries could be evaluated. These dimensions comprise: Power Distance, Individualism vs. Collectivism, Masculinity vs. Femininity, Uncertainty Avoidance, and Long-Term vs. Short-Term Orientation. In developing his model, Hofstede examined Arabic-speaking countries, which he amalgamated together into a single region, based on his assumption that those countries had identical cultural traits. This dissertation seeks to explore Hofstede’s Model by analyzing Arabic websites with the intent to establish whether his regional model is in fact relevant to individual Arab countries. The research used systematic content analysis of 320 websites from 16 Arab countries, including the countries Hofstede originally used in developing his model. Examination of these sites focused on web design elements which have been proven to be prevalent and therefore good indicators of preferences within a particular cultural group. The results showed that while these websites reflect cultural characteristics as presented in Hofstede’s model, they also possessed individual differences, and did not uniformly reflect the design characteristics inferred from the model. These results suggest that Hofstede’s model of cultural dimensions does not fully predict the design of Arabic web interfaces, whether these Arab countries are treated individually or as one group in a region. This research has theoretical implications for the application of Hofstede’s model in usability research, in addition to ii

methodological and practical implications for the localization and design of Arabic web interfaces.

iii

Résumé La culture est l'un des attributs qui affectent l'utilité et l'utilisabilité des sites Web. En intégrant des caractéristiques de conception adaptées à la culture, la localisation d'un site Web qui intègre ces caractéristiques rend ce site à la fois plus attrayant et plus fonctionnel pour ses utilisateurs. La majorité de la recherche dans ce domaine a utilisé le modèle culturel influent de Geert Hofstede. Au moyen de ce modèle basé sur des données recueillies auprès de plus de 50 pays et régions, Hofstede a proposé un modèle composé de cinq dimensions de la culture nationale par lesquels ces pays pourraient être évalués. Ces dimensions comprennent: distance par rapport au pouvoir, individualisme contre collectivisme, masculinité contre féminité, évitement-incertitude, et l'orientation à long terme contre celle à court terme. Quand il a élaboré son modèle, Hofstede a examiné les pays arabophones, qu'il a traités comme une seule région sur la base de son hypothèse que ces pays ont des traits culturels identiques. Cette thèse vise à explorer le modèle de Hofstede en analysant les sites Web arabes avec le but d'établir si son modèle régional est en fait pertinent pour les pays arabes pris un par un. Cette recherche a utilisé systématiquement l'analyse de contenu de 320 sites de 16 pays arabes, y compris les pays que Hofstede a utilisés originellement dans le développement de son modèle. L'examen de ces sites a porté sur des éléments de conception de sites Web qui ont été reconnus comme étant répandus et possiblement préférés dans un groupe culturel particulier. Bien que ces sites reflètent les caractéristiques culturelles tel que présentées dans le modèle de Hofstede, les résultats ont démontré qu'ils possédaient aussi des différences individuelles, et donc ne reflètent pas uniformément les caractéristiques de conception inférées à partir du modèle. Ces résultats suggèrent que le modèle de dimensions culturelles de Hofstede ne iv

prédit pas complètement la conception des interfaces web arabes, que ces pays arabes soient traités individuellement ou en tant que groupe dans une région. Cette recherche a des implications théoriques sur l'application du modèle de Hofstede dans la recherche sur la facilité d'utilisation, en plus d'implications méthodologiques et pratiques pour la localisation et la conception des interfaces Web arabes.

v

‫مستخلص‬ ‫قد يعتبر البعض أن الثقافة هي أحد أهم العوامل التي تلعب دورا مهما فيما يتعلق باستخدام المواقع اإللكترونية ومدى االستفادة‬ ‫منها‪ .‬بناء عليه‪ ،‬فإن تصميم الموقع اإللكتروني بطريقة توافق الثقافة المحلية من خالل إدخال بعض خصائص المجتمع فيه‪،‬‬ ‫يزيد من إقبال المستخدمين عليه ويرفع من فعالية استخدامهم له‪ .‬معظم األبحاث التي تطرقت إلى هذه النقطة‪ ،‬استندت إلى‬ ‫النموذج الثقافي الرائد والمهم للبروفيسور غييرت هوفستد‪ .‬اقترح هوفستد نموذجا مبنيا على أساس معلومات جمعها من أكثر‬ ‫من ‪ 05‬دولة ومنطقة وهذا النموذج يتكون من خمسة أبعاد للثقافة القومية التي يمكن من خاللها تقييم كل بلد على حدى‪ .‬تتضمن‬ ‫هذه األبعاد كال من‪ :‬مقياس فارق القوة بين فصائل المجتمع‪ ،‬والفردية مقابل الجماعية‪ ،‬والذكورة مقابل األنوثة‪ ،‬واالبتعاد عن‬ ‫المجهول‪ ،‬والتوجه طويل المدى مقابل التوجه قصير المدى‪ .‬خالل عملية تطويره لهذا النموذج‪ ،‬قام هوفستد بدراسة بعض الدول‬ ‫الناطقة بالعربية ولكنه اختزلها جميعها بمنطقة واحدة من منطلق أن كل تلك الدول كانت متشابهة تماما في عاداتها المجتمعية‪.‬‬ ‫تهدف هذه الرسالة إلى التعمق أكثر بنموذج هوفستد من خالل تحليل مواقع إلكترونية عربية بغرض إثبات فيما إذا كان نموذجه‬ ‫العالمي ينطبق فعليا على البلدان العربية منفصلة وليست كوحدة واحدة‪ .‬تم االعتماد في هذا البحث على أسلوب التحليل النظامي‬ ‫للمحتوى من خالل دراسة ‪ 025‬موقع إلكتروني لـ ‪ 61‬بلدا عربيا مختلفا بما فيها البلدان التي تضمنها بحث هوفستد األساسي‬ ‫لتطوير نموذجه‪ .‬تركز دراسة هذه المواقع على عناصر التصميم الغرافيكي اإللكتروني والتي أثبت أنها ذات صبغة سائدة بل‬ ‫يطغى عليها طابع التفضيل لدى مجموعة معينة من المجتمع‪.‬‬ ‫أظهرت النتائج بعد البحث والتحليل أنه وبينما قد تعكس هذه المواقع اإللكترونية خصائص اجتماعية كما أوضحها هوفستد إال‬ ‫أنه كانت هناك بعض االختالفات الفردية بين الدول العربية‪ ،‬وعليه فإنها ال تعكس منظومة واحدة لخصائص التصميم التي اشار‬ ‫إليها هوفستد‪ .‬وانطالقا من هذه النتائج فإن نموذج هوفستد لألبعاد االجتماعية ال يمكنه التنبؤ بتصاميم المواقع العربية وسماتها‬ ‫تماما سواء أخذنا البلدان العربية منفردة أو اعتبرناها منطقة واحدة‪ .‬يحتوي هذا البحث على تضمينات نظرية مدروسة لتطبيق‬ ‫نموذج هوفستد في دراسة صالحية استخدام المواقع‪ ،‬باإلضافة إلى اقتراحات منهجية وعملية إلضفاء الخصوصية المحلية على‬ ‫تصاميم المواقع العربية‪.‬‬

‫‪vi‬‬

Dedication ‫إلى أبي وأمي‬ ‫إلى عائلتي‬ ‫إلى نائل‬ To my Dad and Mom To my family To Nael

vii

Acknowledgement I have been blessed with many inspirational people who have touched my life and motivated me throughout my doctoral studies. It is hard to mention all of their names, but they all have my deepest appreciation.

First and foremost, I express my gratitude to my chief supervisor and mentor Professor Andrew Large, who have stood shoulder to shoulder with me during this journey. This work would not have been possible without his knowledge, encouragement, patience, and kindness. I am grateful for the opportunities I had to learn from Andy, whether as a student or as a research assistant. Not to forget that I was lucky for having him as a “real football” fan in a hockey town.

I would like to thank my co-supervisor Professor Jamshid Beheshti for his stimulating discussions, insights, and support, especially in the final year of my doctoral studies. I would like also to thank my committee members Professors Catherine Guastavino, Claudia Mitchell, and Michael Hoechsmann for the invaluable comments and feedback they provided to improve this research.

McGill’s School of Information Studies has provided me with various teaching and teaching assistantship opportunities which will significantly benefit my academic and professional career in the years to come. For its faculty, especially Professor France Bouthillier, I am grateful for their advice and encouragement over the years. My heartfelt thanks go to Professor Elaine Ménard who not only provided me with an exceptional research experience, but also was an amazing boss who has become a great friend. For the staff at the School, Susann Allnutt,

viii

Kathryn Hubbard, Ancy Joseph, and Cathy Venetico, thanks for always being there to help. My sincere thanks go also to Donna Kuzmarov.

Thanks to my doctoral cohort and comrades-in-arms at SIS, for sharing the ups and downs throughout the years. Special thanks to Rhiannon Gainor, Irene Kitimbo, and Jillian Tomm for being great friends and for their peer support that is much needed in our doctoral program. I would also like to sincerely thank Rhiannon for the endless times she proofread my work despite her busy schedule.

Thanks to all my friends who have showed their unconditional support through the good times and the bad, who were excited for my final manuscript submission as much as they were for the first draft, and who cheered me on throughout this course until the finish line. Special thanks go to Hanadi Bani Hani, Emily McHugh, Becky Burbank, and Vilelmini Tsagkaraki. This work was proofread by these friends, as well as Rhiannon, Irene, and Jill, and for this I sincerely thank them all. I would also thank Jerry Fielden and Nadia Jaradeh for translating the abstract into French and Arabic respectively.

Last but not least, my sincere thanks to my family, who knew how important this was to me and therefore has supported me as I followed my dreams. Without the unconditional love and encouragement from each and every one of them, my achievements would not have been realized. Thanks to Nael for teaching me the need to face life difficulties with a smile, his beautiful spirit will always be with me. And no “thank you” is enough to my parents who have given me the confidence that I am capable of doing anything I put my mind to, and I only hope to have made them proud by achieving their dream of having a doctor in the family. ix

Table of Contents Abstract …………………………………………………………………………….…....….….. ii Résumé ………………………………………………………………………………...……….. iv ‫…………………………………………………………………………… مستخلص‬.…...……….. vi Dedication ……………………………………………………………………..……….……….vii Acknowledgement ………………………………………………….……………….………....viii Chapters …………………………………………………………………………...…….……. ..x List of Figures ………………………………………………………………….……….…….. xv List of Tables ……………………………………………………………………………….... xvi

1. Chapter 1: Introduction ............................................................................................ 1 1.1. Background .......................................................................................................................... 1 1.2. Research Problem ................................................................................................................ 5 1.3. Research Questions .............................................................................................................. 7 1.4. Purpose and Significance of the Research ........................................................................... 8 1.5. Organization of the Dissertation ........................................................................................ 11

2. Chapter 2: Usability of Interfaces .......................................................................... 13 2.1. Background ........................................................................................................................ 13 2.2. Usability Principles and User Expectations ....................................................................... 17 2.3. User Interfaces and the Accommodation of Human Diversity .......................................... 18 2.4. Summary ............................................................................................................................ 20 3.

Chapter 3: Culture ............................................................................................................. 21 3.1. Definition ........................................................................................................................... 21 x

3.2. Review of Cultural Models ................................................................................................ 23 3.3. Arab Countries ................................................................................................................... 27 3.3.1. History ............................................................................................................................ 29 3.3.2. Religion .......................................................................................................................... 30 3.3.3. Language ........................................................................................................................ 31 3.3.4. Resources and Wealth .................................................................................................... 32 3.3.5. Governance..................................................................................................................... 33 3.3.6. Modernization ................................................................................................................ 34 3.3.7. Information and Communication Technology ............................................................... 35 3.4. Summary ............................................................................................................................ 38

4. Chapter 4: Hofstede’s Model of Cultural Dimensions ......................................... 41 4.1. The Model .......................................................................................................................... 41 4.1.1. Arab Countries in Hofstede’s Model ............................................................................. 50 4.2. Relationship between the Dimensions and Web Interface Design .................................... 52 4.2.1. Power Distance ............................................................................................................... 53 4.2.2. Uncertainty Avoidance ................................................................................................... 53 4.2.3. Individualism/Collectivism ............................................................................................ 54 4.2.4. Masculinity/Femininity .................................................................................................. 55 4.2.5. Long-/Short-Term Orientation ....................................................................................... 55 4.3. Cultural Influence on Interface Design .............................................................................. 56 4.3.1. Studies of Cultural Influence on Interface Design ......................................................... 59 4.3.2. Studies of Cultural Influence on Arabic Interface Design ............................................. 64

5. Chapter 5: Methodology ......................................................................................... 75 xi

5.1. Method ............................................................................................................................... 75 5.2. Country Selection............................................................................................................... 77 5.3. Genre and Website Selection ............................................................................................. 78 5.4. Unit of Analysis ................................................................................................................. 84 5.5. Unites of Observation ........................................................................................................ 85 5.6. Data Collection Procedures................................................................................................ 91 5.7. Statistical Analysis Procedures .......................................................................................... 93 5.8. Methodological Considerations ......................................................................................... 97 5.8.1. Standardization ............................................................................................................... 97 5.8.2. Data Quality Assurance .................................................................................................. 98 5.9. Limitations ....................................................................................................................... 106 6.

Chapter 6: Results ............................................................................................................ 108 6.1. Power Distance ................................................................................................................ 108 6.1.1. Social Models ............................................................................................................... 109 6.1.2. Status of People in Images ........................................................................................... 110 6.1.3. Symmetry ..................................................................................................................... 111 6.1.4. Images of Buildings ..................................................................................................... 112 6.1.5. Logos ............................................................................................................................ 113 6.1.6. Matching Hofstede’s Score .......................................................................................... 113 6.2. Uncertainty Avoidance .................................................................................................... 116 6.2.1. Menu Structure ............................................................................................................. 116 6.2.2. Number of Link ............................................................................................................ 118 6.2.3. Presence of News ......................................................................................................... 119

xii

6.2.4. Matching Hofstede’s Score .......................................................................................... 120 6.3. Individualism/Collectivism .............................................................................................. 121 6.3.1. Number of People in Images ........................................................................................ 121 6.3.2. Privacy Policy and Rights Reserved Statements .......................................................... 123 6.3.3. Site Registration ........................................................................................................... 124 6.3.4. Site Customization ....................................................................................................... 125 6.3.5. Matching Hofstede’s Score .......................................................................................... 126 6.4. Masculinity/Femininity .................................................................................................... 130 6.4.1. Gender in Images ......................................................................................................... 130 6.4.2. Animated Images.......................................................................................................... 131 6.4.3. Visitors Counter ........................................................................................................... 133 6.4.4. Matching Hofstede’s Score .......................................................................................... 134 6.5. Long-/Short-Term Orientation ......................................................................................... 137 6.5.1. Search Engines and Site Maps ..................................................................................... 137 6.5.2. Help and Frequently Asked Questions ......................................................................... 138 6.5.3. Colors ........................................................................................................................... 139 6.5.4. Matching Hofstede’s Score .......................................................................................... 139 6.6. Design Elements not Associated with Dimensions ......................................................... 142 6.6.1. Entry Page .................................................................................................................... 142 6.6.2. Pictures and Images ...................................................................................................... 143 6.6.3. Country Domain ........................................................................................................... 147 6.6.4. Linguistic Versions ...................................................................................................... 148 6.6.5. Page Orientation ........................................................................................................... 149

xiii

7. Chapter 7: Discussion ............................................................................................ 150 7.1. Power Distance ................................................................................................................ 152 7.1.1. Matching Hofstede’s Score .......................................................................................... 155 7.2. Uncertainty Avoidance .................................................................................................... 156 7.2.1. Matching Hofstede’s Score .......................................................................................... 158 7.3. Individualism/Collectivism .............................................................................................. 160 7.3.1. Matching Hofstede’s Score .......................................................................................... 162 7.4. Masculinity/Femininity .................................................................................................... 163 7.4.1. Matching Hofstede’s Score .......................................................................................... 165 7.5. Long-/Short-Term Orientation ......................................................................................... 166 7.5.1. Matching Hofstede’s Score .......................................................................................... 170 7.6. Other Design Elements .................................................................................................... 171 7.6.1. Entry Page .................................................................................................................... 171 7.6.2. Pictures and Images ...................................................................................................... 172 7.6.3. Linguistic Versions ...................................................................................................... 174 7.6.4. Page Orientation ........................................................................................................... 174

8. Chapter 8: Conclusion.................................................................................................. 177 8.1. Implications and Contributions ........................................................................................ 182 8.2. Future Studies .................................................................................................................. 187

9. Bibliography ................................................................................................................... 191 10. Appendices .................................................................................................................... 208

xiv

List of Figures Figure 1. Arab internet users ...................................................................................................................... 37 Figure 2. Hofstede’s model of cultural dimensions .................................................................................... 45 Figure 3. Code sheet for the Ministry of National Education, Morocco ..................................................... 86 Figure 4. File organizations ........................................................................................................................ 92 Figure 5. Visual map for the results of Kolmogorov-Smirnov test............................................................. 96 Figure 6. A snapshot of the SPSS file for the pilot study ......................................................................... 101 Figure 7. Number of websites with social models .................................................................................... 109 Figure 8. Status of people in images on websites of individual Arab countries ....................................... 110 Figure 9. Number of buildings in images on websites of individual Arab countries ................................ 112 Figure 10. Cluster analysis for design elements related to Power Distance ............................................. 116 Figure 11. Number of websites with simple and complex menus ............................................................ 117 Figure 12. Number of links for each country ............................................................................................ 119 Figure 13. Cluster analysis for design elements related to Uncertainty Avoidance ................................. 121 Figure 14. Number of people in images on websites of individual Arab countries .................................. 122 Figure 15. Number of websites having rights reserved and privacy policy statements ............................ 123 Figure 16. Number of websites with site registration requirements ......................................................... 125 Figure 17. Cluster analysis for design elements related to Individualism/Collectivism ........................... 129 Figure 18. Gender of people in images on websites of individual Arab countries ................................... 131 Figure 19. Number of websites with animated images ............................................................................. 132 Figure 20. Animated images for each country .......................................................................................... 133 Figure 21. Number of websites with visitor counter ................................................................................. 134 Figure 22. Cluster analysis for design elements related to Masculinity/Femininity ................................. 136 Figure 23. Number of websites with site searching tools ......................................................................... 138 Figure 24. Cluster analysis for design elements related to Long-/Short-term Orientation ....................... 142 Figure 25. Number of websites with entry pages...................................................................................... 143 Figure 26. Number of pictures and images from each country ................................................................. 145 Figure 27. Other images for each country................................................................................................. 147 Figure 28. Social models in websites from Jordan ................................................................................... 153 Figure 29. A and B. Hofstede’s Power Distance dimension and cluster analysis for Arab countries ...... 156 Figure 30. 3-level complex menu on Muscat College website, Oman http://www.mctcollege.com ........ 158 Figure 31. A and B. Uncertainty Avoidance dimension and cluster analysis for Arab countries ............. 159 Figure 32. Customization options on the Ministry of Labor website, Bahrain ......................................... 162 Figure 33. A and B. Individualism dimension and cluster analysis for Arab countries ............................ 163 Figure 34. A and B. Masculinity dimension and cluster analysis for Arab countries ............................... 166 Figure 35. A and B. STO dimension and cluster analysis for Arab countries .......................................... 171 Figure 36. Entry page of the Ministry of Planning and International Cooperation website, .................... 172 Figure 37. Images of people appearing while hovering over links ........................................................... 173 Figure 38. Countries scores on Hofstede’s dimensions ............................................................................ 182 Figure 39. FedEx website for Saudi Arabia .............................................................................................. 187

xv

List of Tables Table 1. Members of the League of Arab States ......................................................................................... 28 Table 2. Index scores and ranks for countries and regions (Hofstede, 2001, p. 500) ................................. 43 Table 3. Summary of studies of cultural influence on interface design ...................................................... 64 Table 4. Summary of studies of cultural influence on Arabic interface design .......................................... 67 Table 5. Web design elements associated with Hofstede’s dimensions ..................................................... 74 Table 6. Government web portals ............................................................................................................... 81 Table 7. Category description – web page .................................................................................................. 88 Table 8. Image description .......................................................................................................................... 90 Table 9. Individual pictures – Categories ................................................................................................... 90 Table 10. Intercoder Reliability Results ................................................................................................... 100 Table 11. Statistical analysis for the design elements ............................................................................... 102 Table 12. Background and dominant colors of government portals web interfaces ................................. 103 Table 13. Statistical analysis for the images ............................................................................................. 103 Table 14. Continuous and Categorical Variables...................................................................................... 105 Table 15. Country scores on Hofstede’s Power Distance Dimension....................................................... 152 Table 16. Country scores on Hofstede’s Uncertainty Avoidance Dimension........................................... 157 Table 17. Country scores on Hofstede’s Individualism/Collectivism Dimension .................................... 160 Table 18. Country scores on Hofstede’s Masculinity/Femininity Dimension .......................................... 164 Table 19. Country scores on the Long-/Short-Term Orientation .............................................................. 167 Table 20. South American countries scores on Hofstede’s dimensions ................................................... 184 Table 21. Revised scores for Arab countries on Hofstede’s cultural dimensions ..................................... 184

xvi

1. Chapter 1: Introduction

1.1. Background In his book Weaving the Web, Tim Berners-Lee described his dream of seeing his creation, the World Wide Web, consisting of two main parts: one that provides a powerful means for collaboration between people where they have immediate and intuitive access to information, and another where computers become capable of analyzing all the data on the web, resulting in a “semantic web” (Berners-Lee & Fischetti, 1999). Nowadays, with almost 2.4 billion Internet users worldwide as of 2012 and more than 630 million websites, 51 million of which were added in 2012 alone (Pingdom, 2013), the web provides people with a gigantic information marketplace, where individuals and organizations exchange information with one another. Numbers such as these demonstrate the realization of the first part of Berners-Lee’s vision. The web is surging to encompass the world, becoming an important and permanent part of history, breaking boundaries by pulling people out of their spaces and geographic locations. The web has evolved to become a de facto medium for international communication, where web-based applications have become a standard cross-platform means to store and retrieve information worldwide (Brinck, Gergle, & Wood, 2002). As the web has continued to grow and expand, so too has the awareness that this medium is impacting the communication habits of people all around the globe (Ingram, 2006), and that the design is central to the effectiveness of websites. Vrontis, Ktoridou, and Melanthiou (2006) describe the importance of having such websites within the commercial realm:

1

A website is the heart of the online marketing program, the focus of all marketing communications, and the single most important means of communication for transportation. For prospects, customers, and partners, a company’s website is often the single most visible and used resource. (p. 127) For most users, using the web extends well beyond the basic use of email, to use of popular sites and services, which provide a platform for exchanging information and ideas between individuals. This has opened the door for a global economy, where the increase of global technical product and technology exchanges all over the world have, in turn, offered profound real and potential impact on social relations, culture, politics, and other social activities, including interpersonal interactions, identity formation, retail trade and commerce, and governance (Röse, 2006; Warf, 2013). As the web has emerged and evolved as a distinct media form in recent years, so has the challenge of investigating it as an object of study by researchers from different social, political and cultural backgrounds (Schneider & Foot, 2004). The challenge stems from two main sources: the exact nature of web content, and the hugely diverse backgrounds of web users. Most early websites looked a lot like printed pages pasted onto a screen and were designed with relatively simple interfaces consisting of links, buttons, menus, text, and simple graphics (Brinck et al., 2002). Today, web design differ much more from print design, as the advent of the web as a global communication infrastructure enables concurrent access to distributed information sources through a wide variety of access devices and media, including hypertext, sophisticated graphics, animation, and so on (Bessell et al., 2002; Stephanidis, Paramythis, & Savidis, 2011). Designers are usually responsible for what the website looks like, but the design involves so much more than that. It needs to be functional, aesthetically pleasing, and well constructed, then tested on 2

real people to determine whether or not it works for the intended users (Benun, 2003; Nielsen, 2000, 2005). Diverse user needs become evident when considering the broad range of user characteristics, as well as the changing nature of human activities, the variety of contexts of use, and the increasing availability and diversification of information. There is no typical web user; information artifacts are used by diverse user groups, including people with different cultural, educational, and employment backgrounds, children and adults of all ages, and people with different types of disabilities. For some groups, using a computer in general, and the web in particular, can be a difficult aspiration for many reasons (McKay, 2008). Therefore, the accommodation of users’ attributes for a set of targeted users is essential in order to enhance the usability and appeal of the interface and thus enable all users to experience success and satisfaction (O’Connell & Murphy, 2007). Usability in general could be defined as “the degree to which people (users) can perform a set of required tasks. It is the product of several, sometimes conflicting design goals” (Brinck, Gergle, & Wood, 2002. p. 2). These usability goals or principles include learnability, efficiency, and satisfaction. Enhancing the usability and usefulness of web pages is generally realized by changing either content, structure, navigation, or the interface depending on the current context, which is described by means of user identity, profile, or location (Bomsdorf, 2003; Lee, Kim, & Han, 2008). Users’ needs, expectations, identity, and location are highly influenced by their cultural background (Shneiderman & Plaisant, 2010). Culture is arguably a huge consideration when designing websites because not everybody reads or understands information in the same way,

3

and culture in particular plays a large role in how we view websites. Even the most basic understanding of this principle is needed before designing sites that may be viewed by people from different cultures (Daniel, Oludele, Baguma, & van der Weide, 2011). What becomes clear is that one medium does not equate with one interface. Instead, the interface designs, interactivity, and content reflect some cultural sensitivity and understanding of the targeted audience (Badre, 2001). The relationship between culture and web design usability has been discussed widely in the literature (e.g. Callahan, 2006; Nielsen, 2000; Marcus & Gould, 2000; Masao & Kumiyo, 1996; Fernandes, 1995). The influence of culture on usability is a controversial issue in the field of Human-Computer Interaction. Those who are in support of accommodating culture into the design of user interfaces do not necessarily agree on the level at which culture should be incorporated into user interface design (Ford & Kotzé, 2005). In order to understand cultural presence on the web, one needs to examine and identify elements used to represent it in interfaces, which according to Barber and Badre (1998), are regular features reflecting culture on the web. These design elements can also be seen in relation to Nielsen’s (2005) and Galitz’s (2002) basic usability principles, which state that these design elements should be simple, aesthetically pleasing, adaptive to users’ needs and perspectives, comprehensible, controllable, and familiar in the concepts and language they use based on users’ existing knowledge. The question is, are all cultures being affected by the use of websites in similar ways, leading to greater cultural homogeneity on the web, or are different cultural groups adopting this particular information technology in different ways that are consistent with their respective cultures? In order to understand user interfaces from different cultures, usability experts have heavily used and cited the influential cultural model of Geert Hofstede (1980, 2001), despite the 4

fact that it was not developed for use in information systems or human-computer interaction research (Komlodi, 2005). Many researchers such as Callahan (2007, 2006), Cyr and TrevorSmith (2004), and Marcus and Gould (1998) have employed Hofstede’s cultural dimensions to identify the similarities and differences between cultures through the examination of design elements (e.g., colors, layout, icons, and images) that have been argued to be culturally determined and have an impact on the usability of an interface. Since these elements enhance the performance and appeal of the web when localized for a target culture (Fernandes, 1995), studies have been undertaken to evaluate interfaces for cultural localization based on the interface design criteria associated with Hofstede’s dimensions.

1.2. Research Problem Little research has been done to investigate computer users from different cultures, but it is argued that a good design for one community of users may be inappropriate for another (Shneiderman, 2000). As a result, designers are regularly called on to consider designs for individual cultures. While the potential audience for any website is international and intercultural, very little is known about how websites are perceived and used by individuals from diverse national and cultural backgrounds, especially in communities and countries whose cultural tendencies differ from those where web technologies originate (Zahedi et al., 2001). In the case of Arab countries, interest in the endorsement of Information and Communication Technologies (ICT) in general, and the web in particular, has been growing primarily as a result of economic and social development factors, and is driven by the aim to transform Arab society into a knowledge based society (Abdallah & Albadri, 2011). However, studies of cross-cultural website design and analysis that include Arab countries are scarce

5

compared to what has been published in and on other countries around the world. Prospective web interface culturalization research in this area, which in the past has relied heavily on Hofstede’s model, is problematic for Arab contexts, due to issues such as Hofstede’s limited sample of seven Arab countries, which he considered to be representative of all 22 Arabicspeaking countries. For example, Callahan (2007) analyzed a total of 20 web interfaces from the group of seven Arab countries included in Hofstede’s model in her study of cross-cultural differences in the design of university websites, despite her acknowledgement that Hofstede’s cluster of these countries into one group is somewhat superficial. Moreover, in another methodological issue, Callahan did not identify the number of websites pertaining to each country. Baack and Singh (2007) included only 21 websites from a group of three Arab countries, adding websites from Bahrain, a country that is excluded from Hofstede’s model and thus does not have a score on any of the cultural dimensions. Baack and Singh justified their methodology on the treatment of Arab countries as one region in Hofstede model, and the low number of websites specific to any one Arab country. Further details on Callahan and Baack and Singh studies are provided in the literature review. Assuming that Arab countries are similar is not to consider them identical. To not acknowledge possible individual differences could potentially present a distorted picture of reality. This distinction between ‘similar’ and ‘identical’ raises questions as to whether Hofstede’s model can be appropriately used to examine interfaces from Arab countries, and whether they indeed are representative of each other. It also raises questions about the remaining Arabic-speaking countries, which were excluded from Hofstede’s group: do they exhibit similarities or differences with the original group of seven examined by Hofstede?

6

1.3. Research Questions In light of the importance of investigating cultural presence on the web, and the reliance upon Hofstede’s model of cultural dimensions in the field of cross-cultural web interface analysis, this dissertation seeks to explore and rectify, if necessary, the assumptions inherent in the model by analyzing 320 websites from Arabic countries. The findings of this study will be used to establish whether his regional model is in fact relevant to each individual Arabic country, and examine the implications for Hofstede's original findings in the context of website design. Therefore, the overarching research question is: Can Hofstede’s cultural model explain and predict Arabic web interface design uniformly across all Arab countries? More specifically, this research seeks to answer the following questions: 1. RQ 1: Do websites from the seven individual Arabic countries included in Hofstede’s model all uniformly reflect the interface design characteristics that can be inferred from his dimensions? 1.1. Are there individual differences across countries suggesting that, at least for interface design, Hofstede’s dimensions need to be refined for each country? 1.2. Does any one or more of the seven countries individually match Hofstede’s dimensions?

2. RQ 2: Do websites from the Arabic countries excluded from the Hofstede’s group uniformly reflect the interface design characteristics that can be inferred from Hofstede’s dimensions? 2.1. Are there individual differences between the websites from these countries in terms of interface design characteristics?

7

2.2. Do websites from one or more of these countries match any of those in the group of seven? Several studies have attempted to match web design characteristics to Hofstede’s dimensions (Callahan, 2006, 2007; Robbins & Stylianou, 2002, 2003, 2010). As discussed earlier, the problem of following the same method for Arab countries lies in the shared group scores they were assigned on Hofstede’s dimensions. Furthermore, only six Arabic countries had been assigned such scores in Hofstede’s Arabic grouping, while the remaining countries had been omitted. Therefore, rather than attempting to correlate the design elements to the cultural dimensions, a cluster analysis method was used to identify the closeness of these countries’ scores to the dimension in question and to each other, including the countries excluded from the model.

1.4. Purpose and Significance of the Research With the web now linking diverse regions and communities across the globe that were previously isolated by time and space, the growth of global communications has increased and intensified the need for learning to communicate successfully with a multitude of diverse cultures (Zahedi, Van Pelt, & Song, 2001). This research contributes to the knowledge of web design perceived and used by individuals from countries whose cultural tendencies differ from those where web technologies originate (Zahedi et al., 2001). Hofstede’s model has served as the theoretical basis for many cross-cultural interface analysis studies. Despite the fact that a few studies have included websites from Arab countries, to date no comprehensive study has examined the applicability of Hofstede’s dimensions exclusively to these countries. Another objective of this research, therefore, is the sharp focus on 8

cultural web design for Arab countries only, including in its research design both those original to Hofstede’s model and those excluded. With the exception of a few studies that evaluated interfaces in different cultures (e.g., Callahan, 2007; Robbins & Stylianou, 2003; Zahir, et al., 2002), studies on cultural differences are often limited to specific geographical regions, and also often conducted from a western point of view (Ahmed, Mouratidis, & Preston, 2008). This is reflected in the studies that have compared web interfaces from Europe or North America to their counterparts from other countries, usually East Asian ones (Dong & Lee, 2008; Cyr & Trevor-Smith, 2004; Fang & Rau, 2003). Therefore, another objective of this research is to expand the range of geographic regions included for cross-cultural interface design research. Regions other than North America, Europe, and East Asia have been generally overlooked or inadequately represented in this research area. This study focuses on Arabic-speaking countries, including in total 16 Arab countries, a number far greater than any previous study. For example, Barber and Badre (1998) selected websites from only Lebanon and Saudi Arabia, while Zahir, et al. (2002) selected websites from only Egypt and Morocco. There is an evident scarcity of research on Arab countries in ICT in general and crosscultural interface design in particular, compared to what is being published on other parts of the world (Abdallah & Albadri, 2011). This suggests the need for a comprehensive study that not only identifies any similarities, but also any differences between Arabic web interfaces based on the design characteristics inferred from Hofstede’s model. By doing so, this research is significant in the literature in terms of questioning the appropriateness of grouping Arab

9

countries together, or even choosing a few countries as a representation of the whole culture in cross-cultural web interface analysis. Consequently, this research aims to enrich Hofstede’s model, and its use in the domain of web interface design, upon which many researchers have based their work. If the results reveal significant differences between individual Arab countries in web design based on Hofstede’s cultural dimensions, this suggest that this group should not be treated as a homogeneous entity in cross-cultural interface design. It will also provide insight as to how to evaluate countries that were excluded from Hofstede’s model and thus do not have scores on his dimensions. This research does not directly address the issue of Arabic website standardization or localization, but it can be considered as a first step to providing helpful information for researchers and web interface designers regarding the best practices to follow when inspecting and designing interfaces targeted for Arab users based on design tendencies in these countries. This is because most studies tend to focus on a limited number of web design elements in the interface, missing opportunities to include multiple design elements. For example, Zahir et al. (2002) only inspected color and layout in national web portals from 26 different countries, while Sun (2001) examined four cultural markers (language, visuals, colors, and page layout) in two multilingual websites. This research, on the other hand, provides a comprehensive list of elements that were investigated and inspected on Arabic websites from different Arab countries, which may have practical application in the short/long-term research for Arabic and non-Arabic website developers working with website localization for Arab internet users. It is important to emphasize, however, that this is not a study that examines whether culturalization of an interface improves its usability; rather the aim is to enrich Hofstede’s model, and through it the work of usability researchers. 10

1.5. Organization of the Dissertation Chapter 2 provides background on the usability of web interfaces by discussing the importance of acknowledging and incorporating human factors and user expectations in the design process. Collecting detailed information about users of a particular website is arguably one of the integral and logical first steps in web interface design, which includes the identification of user’s cultural characteristics that have the potential to significantly affect design parameters. Chapter 3 starts with a definition of the term Culture and then summarizes several cultural models in which culture is constructed as a multi-layered concept characterized by several dimensions. The chapter then moves on to provide a background on Arab countries which includes a look at their history, religion, language, and the modernization process they are going through. These factors are at times considered unifying factors for these countries, but they could be argued to be dividing factors as well. As a continuation of the discussion of culture, Chapter 4 provides an extensive preview of Hofstede’s model, which includes its background, a description of the five dimensions, the critique for and against the model, and a description of the treatment of Arab countries in Hofstede’s model, the focus of this research. Then the chapter moves on to describe the relationship between culture and interface design while providing an extensive examination of studies that have utilized Hofstede’s model in their research. Chapter 5 presents the research methodology, which includes a discussion of the content analysis method used in this research, country selection, domain and website selection, the unit of analysis and units of observation, the data collection procedures, the statistical analysis

11

procedures, and finally the methodological considerations for ensuring high quality data collection and analysis including the codebook, coder training and the pilot study. Chapter 6 presents the results of the systematic content analysis of the 320 homepages from the 16 Arab countries included in this research. The presentation of the results is organized based on Hofstede’s five cultural dimensions and the web design elements associated with them. Chapter 7 provides the answer to the research question, whether Hofstede’s cultural model explains and predicts Arabic web interface design. This is accomplished by inspecting and analyzing Arabic websites to determine whether they do in fact reflect Arabic culture as described in Hofstede’s model of cultural dimensions. Chapter 8 provides the conclusion of this research, including the summary of this research, its implications and calls for future research.

12

2. Chapter 2: Usability of Interfaces

2.1. Background People use the web in a variety of ways and their interaction with it varies according to their motivation, proficiency, needs, and expectations. In order to engineer a successful and satisfactory user experience with a website, issues like why people go to a website, what they expect and intend to accomplish, and how they perceive the interface must be examined and understood (O'Connell & Murphy, 2007). Since interfaces are used as a means to interact with computers, designers generally agree that well-designed user interfaces improve the performance and appeal of the web, helping to convert browsers into customers (Marcus & Gould, 2000). The user interface is defined by Thimbleby (1990) as “an information channel that conveys information between user and computer” (p. 18). It is also defined as “…the dynamic interaction that actually takes place when a product meets the users. More precisely, an interface is that interaction that mediates the relation between the user and a tool explaining which approach is necessary to exploit its functions” (Magnani & Bardone, 2006, p. 1). Whether the user interface is of a mobile communication device, information kiosk, or a website, it needs to focus on designing a flexible environment that has a positive impact on the user in order to convey the system’s messages and achieve desirable results. To create these environments, designers have the responsibility to advocate on behalf of the user, ensuring that the interfaces they design are not just merely exercises in technology but that they assist and guide the user from task to task (Watzman, 2008). This is due to the fact that user interfaces are seen as a virtual interactive meeting point between the user’s mental model and the designer’s system model (O'Connell & Murphy, 2007). O'Connell and Murphy described users’ mental models as 13

psychological representations of the ways in which computers and websites work based on users’ previous experience of dealing with them. On the other hand, designers’ models refer to their knowledge of design principles and usability recommendations, which in turn are incorporated into interactive interfaces that promote users’ success and satisfaction. In order to understand users and how they affect the performance of interfaces, several research and experience-based principles are used in web development, including usability engineering (Thimbleby, 1990). Nielsen and Loranger (2006) described usability as “quality attributes relating to how easy something is to use” (p. xvi). More specifically, it refers to how quickly people can learn to use something, how efficient they are while using it, how memorable it is, how error-free it is, and how much users like using it. This is an important issue to consider because, according to Nielsen and Loranger, users are extremely impatient as they spend an average of only 27 seconds on each web page, moving from one page to another, but certainly spend more time on web pages proven useful in their past experience. Usefulness to users means simple, comprehensible, predictable, and controllable; users feel competent, satisfied, and responsible for their actions (Shneiderman & Plaisant, 2010). Looking specifically at the web environment, there are a number of its characteristics which pose a challenge for human-computer interaction researchers and professionals. These characteristics include a distributed network, a diverse distributed population of users, and various different technologies, browsers, platforms, connection speeds, and plug-ins (Lazar, 2003). Despite the fact that the relationship between the web and usability has been developing since the 1990s (Ashman et al., 2008), usability on the web is often ignored by designers who design for their own pleasure instead of trying to satisfy user needs (Nielsen & Loranger, 2006). Ashman et al. argue that not only does the web demand good usability, but that usability 14

necessitates the web, and that there are a few reasons why these two terms have become so closely associated with each other, such as: 

Many websites have been designed with as few designers’ technical skills as possible (e.g. programming languages such as JavaScript and PHP), which has ultimately added to the complications of navigation



Web users are largely discretionary users, who generally do not have to use a particular site. If they experience usability problems, they do not necessarily have to struggle with or adapt to the poor interface, they are empowered to explore other web options



Web usability problems have a clear relationship with sales, therefore if users cannot find the information they are looking for, they are likely to quit the site



The technical characteristics of the web are continually changing in response to new applications. This has an impact on the functionality available to web users (i.e. animations, videos), developments in user interface design (i.e. clickable items), as well as tools (e.g. plug-ins). In contrast, users’ skills, knowledge, and expectations are considerably slower to evolve, leading to an inevitable gulf between users and the web



Web users are extremely diverse, which raises a large number of issues related to the user variables such as age, gender, culture, disability, language abilities, computer skills, and so on.

The accommodation of users’ attributes for a set of targeted users in particular is essential for the web design process in order to enable users to experience success and satisfaction initially intended by the design process (O’Connell & Murphy, 2007). Therefore, the role of the designer

15

is to facilitate the task for the user and to make sure that the user is able to make use of the website as intended, and with only a minimum effort required in learning how to use it (Abras, Maloney-Krichmar & Preece, 2004). This will require a consideration of the human factors that affect users’ performance, which Shneiderman and Plaisant (2010) summed up as follows: 1. Time to learn 2. Speed of performance 3. Rate of errors by users 4. Retention over time 5. Subjective satisfaction

The purpose of usability engineering is to raise the potential for users’ success and satisfaction and to support website providers’ goals. However, the complex set of variables residing in any user group must first be understood, and then this understanding must be applied to promote users’ success and satisfaction. To take account of users’ needs in terms of interface usability, it has been suggested that users should be involved in the design process (Nielsen, 2000; O'Connell & Murphy, 2007; Shneiderman & Plaisant, 2010). This process requires defining the roles and activities of users, which depend on common definitions of concepts, inputs, outputs, and tools. This understanding supports the collaboration necessary to incorporate user-centred processes into website development (O’Connell & Murphy, 2007). Therefore, if an interface is really considered as an extension of a person, according to Galitz (2002), it should reflect a person’s capabilities and respond to his or her specific needs and match their own expectations.

16

2.2. Usability Principles and User Expectations A user has a lot to expect from the interface he or she is interacting with. As argued by Galitz (2002) and Nielsen (2000), it should be useful, that is, accomplishing user’s objectives fast and efficiently, easy to learn, and enjoyable and fun to use, evoking a sense of pleasure and accomplishment rather than boredom and frustration. Based on users’ expectations, Galitz presented some usability principles which enhance the performance of an interface: 

Aesthetically pleasing and attractive by providing visual appeal to the user, yet consistent in this appeal



Simple, clear, and concise in its metaphors, functions, visual elements, and text



Maintain a compatibility with user and task and therefore adopt the user’s perspective



Comprehensible and easily understood through the flow of actions, responses, visual presentations, and information



Controllable by and responsive to the user’s requests and direct in task accomplishment



Familiar in the concepts and language it uses based on users’ existing knowledge



Flexible to differing needs of users (knowledge and skills, experience, personal preferences, and habits)



Forgiving of unavoidable common errors, yet provides constructive messages in case they occur



Predictable in the way it anticipates the natural progression of a task



Recoverable by having reversible actions



And transparent as working inside computers remains invisible to the user These principles are consistent with Nielsen’s (2005) usability guidelines which include

learnability, utility, efficiency, memorability, error management, and satisfaction. While Brinck et al. (2002) consider these principles as design goals, they argue that they can conflict with one another and therefore the context of the design will determine the priority with which principle is applied. For example, an interface for a video game is expected to be aesthetically pleasing and 17

attractive, while an information kiosk system would probably have simplicity and ease of learning as a primary usability goal. However, these principles nonetheless require the inclusion of users in the interface design process.

2.3. User Interfaces and the Accommodation of Human Diversity The literature that has examined users in relation to interface design is diverse and comes under many disciplines like computing, psychology, anthropology, and linguistics (Thimbleby, 1990). It is as remarkable as the diversity of human abilities, backgrounds, motivations, personalities, and work styles that continuously challenge user interface designers. Thus, designing and building a successful website for the targeted user group means seeking to understand their mind and behavior by acknowledging human tendencies and limitations as well as make allowances for them in their overall design (Benun, 2003; Vergo, Noronha, Kramer, Lechner, & Cofino, 2003). Collecting detailed information about users of a particular website is one of the integral and logical first steps in web interface design. Vergo et al. (2003) noted that from a designer’s perspective, this phase requires the identification of user’s background and characteristics that have the potential to significantly affect design parameters. From a marketer’s perspective, it requires the identification of the groups of customers that are likely to be the most profitable consumers of the products and services that are being offered on the website. Consequently, understanding the physical, intellectual, and personality differences among users is essential for the usability of web interfaces. Shneiderman and Plaisant (2010) listed several factors relating to user diversity that should be taken into account when designing user interfaces: 1. Physical abilities 18

2. Cognitive and perceptual abilities 3. Personality differences 4. Users with disabilities 5. Elderly users/children 6. Cultural and international diversity Brinck et al. (2002) argues that these different types of users will influence the selection of content and functionality for the site, the way text is worded, the appearance of the site, and the level of simplicity of the design. For example, there are a number of settings where elderly people are likely to encounter communication technologies, including the web. However, this user group is at a disadvantage when using computers because they have not had the exposure to them either culturally or through education (Czaja, & Lee, 2008; Gabriel & Benoît, 2003; Noyes & Sheard, 2003; Richards, Hanson, & Trewin, 2003). Therefore, designers need to take into consideration their needs and expectations from a system and its interface, such as having few colors, few images, low density of information, size of text, contrast between text and background, reduced number of links per page, etc. (Gabriel & Benoît, 2003). By taking these elements into consideration for this user group, designers accommodate cognitive and physical aging processes that affect seniors’ performance on and satisfaction with the web (Noyes & Sheard, 2003), thus enabling them to experience success and satisfaction with software to a degree comparable to that enjoyed by people without special needs (O'Connell & Murphy, 2007). However, many other diverse users are still not considered and need to be included when designing interfaces, for example, low income, homeless, and unemployed users (Shneiderman, 2000).

19

With the staggering number of multinational users from around the world, and the global reach of the web (which allows culturally diverse users to access complex data and functions), many researchers asked whether culture actually matters on the web, and if so, how it affects web design and usability (Burgmann, Kitchen, & Williams, 2006; Tsikriktsis, 2002). Marcus (2003, 2008) argues that the global user interface design consists of partially universal and partially local solutions to the design of metaphors, mental models, navigation, appearance, and interaction. Either way, the development of the web user interface must account for the targeted user group’s culture. Little is known about computer users from different cultures, but what is known for sure is that a good design for one community of users may be inappropriate for another, and as a result designers are regularly called on to make designs for individual cultures.

2.4. Summary In order to design and develop successful web user interfaces, designers are called on to consider human factors that affect users’ performance and satisfaction with the web. To some extent, cultural dimensions can be applied to identify differences amongst cultures that may have an effect on website design, such as language, communication style, time dimension, etc.

20

3. Chapter 3: Culture

3.1. Definition Culture is regarded as one of the most complicated words in the English language to define, partly because of its complex historical development and also because it has now come to be used for important concepts in several distinct intellectual disciplines (Bennett & Frow, 2008; Thompson, 1990; Williams, 1958, 1976, 1981). By looking at the work of Kroeber and Kluckhohn (1952), one can apprehend the complexity of culture as a term. Kroeber and Kluckhohn accumulated 156 definitions of culture and classified them under six aggregate headings: descriptive definitions that attempt to identify the content of culture; historical definitions that focus on the social heritage and tradition; normative definitions that focus on the rules and ways of behaving in cultures; structural definitions which have an emphasis on statistical regularities; genetic definitions that characterize culture in terms of products, ideas, and symbols; and psychological definitions which attempt to attach culture to a process of adjustment, learning, and development. This corresponds to what Hall (1959) noted when he said that culture is a word already having so many meanings that one more can do it no harm. Williams’ Keywords: A Vocabulary of Culture and Society (1976) provided an effective and reliable variety of past and present meanings attached to the discussions of culture, a work that has been revised twice (Bennett, Grossberg, & Morris, 2005; Williams, 1983) to reflect shifts in the term’s use. Many researchers react to the complex nature of the term Culture by opting to select one of its definitions, conceptually and operationally define it, and then examine it within a specific discipline, sometimes through concrete measurable elements, to understand and explain why and how cultures differ. 21

Social scientists use the term loosely as a synonym for society or to refer to geographic variations in experience and environment when using it as an independent variable (Fiske, 1996). But for anthropologists, culture can be seen as a technical term used to refer to a system for creating, sending, storing and processing information developed by human beings, which differentiates them from other life forms (Hall & Hall, 1990). Williams (1981) sees culture as the signifying system through which by necessity a social order is communicated, reproduced, experienced and explored. He defined it as “[...] the active cultivation of the human mind, [...] a noun of configuration or generalization of the ‘spirit’ which informed the ‘whole way of life’ of a distinct people” (p. 113). Kroeber and Kluckhohn (1952) dentified culture as “patterns, explicit and implicit, of and for behavior acquired and transmitted by symbols, constituting the distinctive achievement of human groups, including their embodiment in artifacts; the essential core of culture consists of traditional (i.e. historically derived and selected) ideas and especially their attached values; culture systems may, on the one hand, be considered as products of action, on the other as conditioning elements of further action” (p. 357). According to Hoft (1996), it is “the manifestation of learned behavior consisting of thoughts, feelings and actions, formed under the influence of social, biological, psychological, economic and ecological environments over an extended period of time” (p. 41). Hofstede (Hofstede & Hofstede, 2005) considered culture as “the collective programming of the mind that distinguishes the members of one group or category of people from others” (p. 4). Culture is a collective phenomenon because it is at least partly shared with people who live or lived within the same social environment, which is where this culture was learned. Hofstede also highlighted essential patterns of thinking, feeling, and acting that are well established by late 22

childhood. Cultural differences manifest themselves in a culture’s choices of symbols, heroes, rituals, and values where symbols represent the most superficial and values represent the deepest manifestations of culture, with heroes and rituals in between (Hofstede & Hofstede, 2005). Culture is something that comes naturally to societies, as Williams noted (1989), since every society has its own shape, its own purpose and its own meanings which are expressed in institutions, arts, and in learning. Although each of these definitions underlines specific concepts associated with culture, they do share some similar features. Culture is socially created, defined, and managed by a group of people over a long period of time; manifested in behavior, values, patterns of thinking, and communication style; and it is expressed by language, artifacts and symbols. Jones (2007) adds that these ingredients of culture are acquired from birth. They are influenced by family, school, religion, workplace, friends, television, media, and many other sources. Yet, culture is open to interpretation because regional cultures of the world occur both by nation and ethnic group and, more broadly, by larger regional variations. Many regional cultures have been influenced by contact with others through trade, migration, media, and religion; thus culture is dynamic and changes over time (Baskerville, 2003; Mani, 2002; McSweeney, 2002). In doing so, cultures absorb external influences and adjust to changing environments and technologies.

3.2. Review of Cultural Models In an attempt to construct culture as a multi-layered concept, several models have been developed to characterize its dimensions. Hoft (1996) described the following models as metamodels and summarized them to give a deeper understanding the term Culture:

23



The objective and subjective culture model: where objective culture is described as the institutions and artifacts of a culture, for example, its political structures, economic system, social customs, arts, crafts and literary works. Subjective culture encompasses the psychological features of the culture which includes values, assumptions, and patterns of thinking.



The Iceberg model: which depicts culture as an iceberg where only 10% is visible above the surface of the water, which means that only 10% of cultural characteristics are easily visible to the observer, while 90% of the characteristics are hidden from view.



The Onion model: presented by Trompenaars (Trompenaars, 1994; Trompenaars & Hampden-Turner, 1998) symbolizes culture as an onion where the outer layer represents objective culture as the first things we encounter when introduced to a particular culture. The middle layer defines norms and values. And the core represents implicit characteristics of the culture and consists of how people adapt to their environments. This model is similar to the subjective-objective cultural model.



The Pyramid model: was introduced by Hofstede in which he positioned culture (understood as learned, not inherited characteristics common to a specific group or category of people) midway between personality specific to an individual (inherited and learned) and human nature (inherited) common to all human begins. The boundaries between personality and culture, and culture and human nature here are blurred.

Edward Hall defined culture as a “shared communication system within one group or region” by which we can come to recognize and understand the unexplored region of human behavior in our 24

own and other cultures (Hall & Hall, 1990, p. 3). This communication system (i.e. culture) consists of four dimensions: 1. Level of context, which is the most often applied dimension according to Komlodi (2005), divides cultures into two main types: high- and low-context cultures. A high context culture places more emphasis on the unspoken meaning of a given situation than on the actual message. On the other hand, cultures with low context interaction place much more importance on the explicit message. 2. Time dimension; it also has two main types: polychronic and monochronic. Polychronic time is circular, in which many actions and tasks can take place at the same time, while monochronic time is linear in which one event happens at a time. 3. The speed of messages describes the frequency and pace of messages members of various cultures find acceptable. While some cultures are used to faster-paced messages, others prefer slow-paced messages. 4. Finally, the information flow dimension addresses how long it takes a message to travel and produce the desired effect.

Trompenaars described culture as “a shared system of meaning that dictates what we pay attention to, how we act, and what we value”, and also “the way in which a group of people solves problems and reconciles dilemmas” (Trompenaars & Hampden-Turner, 1998). These problems arise from relationships with other people, problems related to time, and those related to environment.

25

According to Trompenaars and Hampden-Turner (1998), relationships with other people are based on five relational orientations: universalism vs. particularism, individualism vs. communitarianism, neutral vs. emotional, specific vs. diffuse, and achievement vs. ascription. While universalistic cultures focus on the social rules, particularistic cultures focus on relationships between individuals. Individualism vs. communitarianism primarily focuses on the individual or group orientation of the culture. Neutral vs. emotional orientation describes the cultural preference towards the expression of emotions. Specific vs. diffuse orientation relates to cultures’ different preferences to separate work from private life. Achievement vs. ascription describes the culture’s emphasis on personal achievement or personal connections. Attitude towards time also differentiates cultures as some appreciate the past while others focus on the future. Additionally, cultures differ in people’s attitudes towards their surrounding environment and whether they think that they control this environment or it controls them. Nisbett’s cognitive model (Masuda & Nisbett, 2001; Nisbett & Norenzayan, 2002; Nisbett, Peng, Choi, & Norenzayan, 2001) differentiates between holistic and analytic thought, especially in the case of Asian and American cultures. Asian culture, which Nisbett described as being holistic, has an orientation tendency to the context as a whole, a preference for predicting events on the basis of the relationship between a focal object and the field, an approach that relies on experience-based knowledge, an emphasis on change, and recognition of contradictions. American culture, which is considered as an analytic thought culture, detaches the object from its context, has a tendency to focus on the attributes of the object, has a preference for using rules about categories to explain an object’s behavior, and uses formal logic and avoidance of contradiction.

26

Since the term Culture could be used to refer to geographic variations or similarities in experience and environment (Fiske, 1996), one can say that Arab countries arguably share the same culture based on the similarities in patterns of living and production as well as value orientations in all Arab countries (Barakat, 1979).

3.3. Arab Countries Arab countries refer to those Arabic-speaking countries that extend their sovereignty over territories and populations distributed over a geographical and ethnic area stretching from the Atlantic coasts of Mauritania and Morocco to the shores of the Persian Gulf, the Indian Ocean, and the African coasts of the Gulf of Aden. Although sometimes these countries are referred to as the Middle East, the latter term includes other countries that are not considered part of the Arab world, such as Iran and Turkey (Gelvin, 2005), as well as omitting Arabic-speaking countries in North and East Africa. Nishio (2001, pp. 137-138) states that the word ‘Arab’ has two meanings for those living in the Arab world, one refers to the nomadic Arabs while the other refers to the Arab race. The former usage has existed since long ago, while the latter has only existed since relatively modern times, and has political origins. However, it must be pointed out that even when the word ‘Arab’ is used with its meaning of the ‘Arab people’ or ‘Arab race’, there are a variety of possible meanings depending on the cultural, historical, political or geographical context, but that the idea that it refers to ‘speakers of Arabic’ is central to its meaning. Al-Sayyid (1973) defined the Arab nation as “that human collectivity which speaks the Arabic language, inhabits a territory called the Arab lands, and has a voluntary and spontaneous feeling of belonging to that nation” (p. 49). The body which tries to coordinate the views and 27

actions of the various Arab countries is the League of Arab States which was formed in 1945 by a handful of Arab countries (Butt, 1987). Based on the League’s website (http://www.lasportal.org), there are currently 22 countries which are members of the League (see Table 1), 21 of which are members of the United Nations (Palestine is excluded1).

Table 1. Members of the League of Arab States Country Egypt Iraq Jordan Lebanon Saudi Arabia Syria Yemen Sudan Libya Morocco Tunisia Kuwait Algeria Bahrain Qatar Oman United Arab Emirates Mauritania Somalia Palestine Djibouti Comoros

1

Joined Arab League in 1945 1945 1945 1945 1945 1945 1945 1956 1956 1958 1958 1961 1962 1971 1971 1971 1971 1973 1974 1976 1977 1993

Palestine was recognized as a non-member observer state by the United Nations on 29 November 2012

28

It could be argued that the identity of these countries is wrapped around several important shared elements that unify the region, such as geography, history, religion, and language (Kamrava, 2005; Matar, 2012).

3.3.1. History The beginning of the history of this region is a matter of some debate. Some researchers, such as Gelvin (2005), support the idea of going back as far as the emergence of Islam in A.D. 622, when the Prophet Muhammad established the first lasting Islamic community in the Arabian Peninsula. Later, much of the Arabian Peninsula joined or was affiliated with the Islamic community, and within a hundred years of the Prophet’s death, Arab/Muslim armies had conquered all of Persia, Mesopotamia, Egypt, and the Levant. They had pushed back the frontier of the Byzantine Empire in the north, had traversed North Africa, and had crossed the Mediterranean. The Islamic conquests were followed by the settlement of Arabs throughout the conquered territories. Other researchers, like Kamrava (2005), favour the deployment of the early 20th century, after the collapse of the Ottoman Empire and in the wake of World War I, as a starting point for the modern history of the region. Much of the Arab world, with the exception of Morocco, experienced centuries of Ottoman rule, generally from the mid-16th century up until the late 19th century. Once this empire collapsed, the European powers took its place, leaving their own distinctive marks on the region. According to Kamrava (2005), the decisions made by these powers, motivated by political or strategic considerations, had little to do with the cultural, geographic, historical or other realities of the region. The most important of these decisions involved the division of the Arabic-speaking portions of the former Ottoman Empire into virtual

29

colonies of European countries. Britain received the mandate for the territory that now includes Israel, the occupied territories, Jordan, Iraq, and Egypt (Gelvin, 2005). Goldschmidt (2002) added that despite the fact that most of the Arabian Peninsula was not under the Ottoman rule, Britain drafted treaties that would have made this land virtual British protectorates, as in the cases of Kuwait, Bahrain, the Trucial States (now called the United Arab Emirates), and Oman. France, on the other hand, received the mandate for the territory that now includes Syria, Lebanon, Algeria, Tunisia, and part of Morocco while the other part of Morocco was occupied by Spain, and Italy occupied Libya (Findlay, 1994).

3.3.2. Religion Almost 95% of the 300 million Arabs are Muslims, making up 20% of the 1.57 billion Muslims around the world (Nydell, 2006). They share a compelling set of beliefs and rituals that transcend national boundaries with remarkable ease (Kamrava, 2005). According to Choueiri (2000), Islam turned the Arabs from being an ethnically-based community into a nation sharing a literary and ethnic movement that adapted the myths, memories, symbols and values of Arab civilization to new conditions, endowing them with new meanings and new functions, which constitute the Arabic language and historical Islam as the cornerstones of the nation. Christians and Jews are considered minorities in the Arab world. While Christians are concentrated in Lebanon, Syria, Jordan, Egypt and Palestine, most Arab Jews left or were expelled from their countries of birth from the late 1940’s to the early 1960’s, following the creation of the state of Israel, and are now mostly concentrated in Israel. The few remaining Jews in the Arab countries reside mostly in Morocco and Tunisia (Findlay, 1994).

30

3.3.3. Language Arabic is not only the mother tongue of millions of people and is the most widely spoken language in the Arab world, it is also considered as one of the pillars of the Arab nation (AlSayyid, 1973) and an important indicator of the Arab identity (Nishio, 2001). Although the language existed long before Islam emerged as a religion in the region in A.D. 622, it attained the status it has today because of Islam (Choueiri, 2000). Arabic remains the language originally chosen as the vehicle of the sacred and the mirror in which the whole of Islam contemplates itself (Messadi, 1981). Besides being the language of the Quran, Nishio (2001) notes that Arabic had another two symbolic functions that added to its advancement: being the language of the scholars and the communication medium for Islamic civilization used by the political and cultural elites. However, the language situation changed under the Ottoman administration after the 16th century where the role of Arabic in the culture, but not in religious activities, diminished to a great extent and was replaced by Turkish. After the collapse of the Ottoman Empire, the European powers took its place, leaving their own distinctive marks on the language. In countries that were under the British mandate such as Egypt, Jordan, and Iraq, English remains strong even today. The same situation could be said for countries which were heavily affected by the French colonization policy such as Algeria, Tunisia, Morocco and Lebanon, where the influence of French also remains strong. Ibn Al-Faruq (2002) provides an example of the language situation in Morocco. He notes that despite the fact that Arabic is the official language of the state in Morocco, it actually remains weak in terms of its functional status in the society because it is not used in daily communication. French, on the other hand, occupies a distinct status there, as it is taught from 31

the early stages of education till the end of university education. It is also used in different administrative, economic, and cultural activities, and thus considered to be in constant competition with Arabic. There is also the competition from languages spoken by the minorities in the Arab world, such as the Berber language spoken either as a first or second language in Algeria, Morocco and Libya (Bengio & Ben-Dor, 1999). The influence of foreign and indigenous local languages led some Arab countries to adopt several official languages besides Arabic. The Factbook (2012) states that Comoros has three official languages: Arabic, French, and Shikomoro. Djibouti has two official languages: Arabic and French. On the other hand, Hassaniya is the official language of Mauritania, while Somali is the official language of Somalia. Gill (1999) pointed out that on an individual level, language choice in the Arab world is governed by the need or the will to adapt to socio-economic circumstances, where demand tends to be motivated by notions of economic gain and social prestige. Therefore, according to Gill, it is possible to see English seriously challenge the position of French in North African countries as a vehicle of Western communication in a fairly distant future. For this reason and giving the example of Moroccan society, Ibn Al-Faruq (2002) calls for the need of a policy that would strengthen and enforce the use of the Arabic language through supporting it by all possible means in the educational system, and putting a focus on the Arabization of institutions in the country.

3.3.4. Resources and Wealth The most striking feature of the region is the vast deserts that span many countries, where there are no grassy plains and more than nine-tenth of the forests have been cut down 32

(Goldschmidt, 2002). Nonetheless, Arab perspectives of the region are dominated, not by the deserts, but by the water resources made available by great rivers such as the Nile, Euphrates and Tigris which bring water into the area and have been the focus of settlement patterns throughout most of Arab history, and also the reason for disputes due to water scarceness (Findlay, 1994). Some Arab countries such as Morocco, Algeria, and Jordan harbour deposits of copper, iron, and other metals, however, oil remains the most notable natural resource of the area. Nevertheless, not all countries have this asset as only Saudi Arabia, Iraq, Kuwait, and the United Arab Emirates have immense petroleum deposits, with much lower reserves in Oman, Qatar, Egypt, and Algeria (Goldschmidt, 2002). The income generated by oil production has provided the Arab oil states with massive capital reserves to deploy in relation to their development plans. Libya, for example, devoted very significant portions of its oil revenues in the 1980’s to building a man-made river to pump water from underground reservoirs in the south of the country to the coastal agricultural regions. Oil has also given countries such as Saudi Arabia and the United Arab Emirates the power and prestige that challenge the position of other leading countries like Egypt and Syria, which historically have been the cultural and political capitals of the Arabs (Findlay, 1994).

3.3.5. Governance Besides the artificial national borders that were established, the British rule left the institution of monarchy as its biggest relic, which was secured in almost all the lands Britain ruled, from Egypt to Jordan, Iraq, and the Arabian Peninsula. On the other hand, the French left behind republican systems that mimicked their own, and also ensured the prosperity of their

33

civilization by making French the informal language of the North African countries except for Egypt (Kamrava, 2005).

3.3.6. Modernization There is an immense modernization process currently taking place in the Arab world, where social change is evident everywhere due to the adoption of economic innovation, political experimentation, educational concepts, and the implementation of Western technology (Nydell, 2006). Taking the advancement of educational systems as an example, Nydell noted that overall rates of literacy have skyrocketed in the Arab world since the 1960s. In the last 40 years the number of educated people doubled in some countries and increased ten times or more in others. Literacy in the Arab countries is at an average of 68%, and among those living in urban areas, it has reached 80% or more almost everywhere. The highest rates of literacy are in Saudi Arabia (94%), Jordan (91%), Bahrain (89%), Lebanon (86%), Qatar (85%), and Kuwait (83%). The rise in literacy is due to a phenomenal growth in public education, which is free and mandatory in all Arab countries (Nydell, 2006). Despite the fact that modernization and globalization have been inevitable and are present on different levels, they have raised controversies over their implementation in Arab countries (Nydell, 2006). According to Hammond (2005), fear about the effect of globalization and what foreigners and their cultures may have in store for the Arab culture is currently widespread in the Arab world, to different degrees in the individual countries. In some countries, like Syria, the reasons for the slow adaptation of modernization are political, while for poor countries like Yemen, they are economic. While the Arab identity has gelled, fears of its disintegration are high, and many Arab intellectuals and politicians wonder whether the Arabic countries are destined to 34

become nothing more than a group of nations that happen to speak a similar language. Hammond (2005) presents another point of view by stating that modern media and communication have increased the volume of the cultural reaffirmation that has transpired among countries of the region and also between individual countries and the rest of the world, emphasizing similarity, homogeneity, and a greater sense of Arabness or Arab identity.

3.3.7. Information and Communication Technology The globalization of media has had a strong influence on the development, policy, and regulation of information and communication systems all over the world, including Arab countries (Eid, 2009). Interest in this growing endorsement of information and communication technology (ICT) is primarily attributed to economic and social development factors, and is driven by the aim to transform Arab society into a knowledge based society (Abdallah & Albadri, 2011). The level of the digital divide, however, and the progress of adopting ICT in Arab countries vary significantly at the individual level between these countries. Hamade (2009) and Allagui (2009) accredit this to infrastructure and economic problems (electric power shortage, extreme poverty, and high internet costs) and also on policy and regulation problems (social and legal constraints and ICT skills shortage). Al-Mabrouk and Soar (2009), on the other hand, state that the characteristics, environment, government, and lifestyles in the Arab countries are different and the culture itself is different from other cultures, which may not help IT transfer in these countries. According Al-Mabrouk and Soar, the present pace of transfer and use of IT is far behind that found in many other developing countries that are in the same stage of development, like Asian and Latin American countries. Fletcher (2006) pointed that for the web in particular, adoption varies due to technology and reasons related to level of development, including the lack 35

of necessary infrastructure, the dominance of English as the language used on the web, and the dependence of the medium on interactivity. However, Fletcher also pointed out that technology adoption is due to cultural factors such as reluctance to innovate, a focus on the past as opposed to the present or the future, collectivism, uncertainty avoidance, and the high degree of dependency on context. Figure 1 provides a comparison between Arab countries in terms of percentage of Internet users based on 2012 estimates (Internet World Stats, 2012). As can be seen in the figure, the Gulf States, including Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and the United Arab Emirates (UAE) have the highest numbers of Internet users (ratio of users to the population in each country). This is contributed to attracting ICT investment due to the rich oil economy (Barry, 2011; Hamade, 2009; Mezher, 2007). Interestingly, Abdallah and Albadri (2011) note that Saudi Arabia, although considered a wealthy country, had a slower ICT adoption and still have a penetration lower than other Gulf States due to the conservative nature of the Saudi culture and its scepticism about the role of technology and its impact on a traditionally religious society. On the other hand, the poorest countries, like Mauritania, Somalia and Yemen, have the lowest penetration rates, while Iraq’s penetration rate is very low due to the existing turmoil.

36

Figure 1. Arab Internet users http://www.internetworldstats.com/ Allagui (2009) adds that despite the fact that Lebanon and Tunisia were among the first Arab countries to connect to the Internet, the war in Lebanon destroyed the infrastructure and drained the country into a digital hole. Whereas in Tunisia, a lack of positive public policy towards a wider and more expanded access to the Internet made the country close to being among the Arab laggards. Interestingly, Jordan and Lebanon have only moderate economies and

37

government support for ICT, yet still their penetration is relatively high, perhaps attributable to the skilled workforce and openness of their markets (Mezher, 2007). Some of the countries that have high penetration rates also have well-established practices in using ICT in different domains. The application of ICT in education is prominent in Arab countries in order to reduce the digital divide. Countries such as the UAE, Egypt, Jordan, Lebanon, Qatar and Saudi Arabia have made extensive use of ICT in schools and universities, and ICT literacy programs are well integrated into education at the primary, secondary and higher education levels (Abdallah & Albadri, 2011). At the government level, various applications and implantations have taken place in several countries to provide government services for citizens through web portals that include paying taxes and renewal of passports, among other services. Again, most of the Arab countries that are advanced in this domain are those of the Gulf States or those that have open markets such as the UAE, Jordan and Bahrain, which are ranked among the top 50 countries on a global scale rating (Abdallah & Albadri, 2011). Loch, Straub, and Kamel (2003) noted that cultural obstacles to the diffusion of information technologies in the Arab world are significant, but there are many factors that are encouraging as well. They argued that if cultural beliefs and attitudes toward technology were better understood, then the technology itself, and more specifically the Internet, might be better adapted to the behavioral patterns of each country, rather than the traditional approach of forcefitting the culture to the technology.

3.4. Summary Intentionally or otherwise, there are trends that treat Arabs as a whole, while other trends treat them as smaller units (Hammond, 2005). From one perspective, Arabs share a common 38

heritage of customs, collective feelings, language, and social and religious views. These elements, paired with other factors such as sharing cultural, geographical and economic realities, as well as sharing common historical experiences such as colonization and the struggle for independence constitute an integral part of Arab unity (Barakat, 1979). The other perspective argues that these same unifying factors are so varied, and sometimes even so dissimilar, that there is a temptation to consider these countries not as a composite and coherent whole, but as an assembly of heterogeneous elements merely placed side by side (Butt, 1987; Kamrava, 2005; Messadi, 1981). This can be seen in Zogby’s (2010) argument that each country has its own unique situation where Cairo (Egypt) is not Riyadh (Saudi Arabia), is not Beirut (Lebanon), is not Marrakesh (Morocco), despite that this region is frequently portrayed as one homogenous unit by commentators, politicians, or academics. Hofstede and Hofstede defined Culture as the “the collective programming of the mind distinguishing the members of one group or category of people from another" (n.d.), where "category" can refer to nations or regions within/across nations, among other possibilities. Fiske (1996) also noted that Culture could be used to refer to geographic variations or similarities in experience and environment. For the purpose of this research, Culture refers to the similarities and differences in experience and environment at the national level between Arab countries, where the focus is on the objective (visible) variation. Zayani (2012) notes that there is a vacuum of developed theoretical and conceptual models that take into account the specifity and complexity of Arab countries, whether being developed in the region or elsewhere. In the absence of such models, Western academic discourse on Arab countries has been dominated by theories and conceptual models which were produced outside the region and around specific institutions, technologies, and political systems. 39

Zayani argues that these theories, when applied indiscriminately and uncritically to the study of Arab media in particular, are problematic because they are based, in large part, on a set of unrevised assumptions which are insufficient to explore the dynamics of the region. This can be seen in the case of the influential cultural model developed by Geert Hofstede (1980, 2001), where Arab countries were represented by only seven countries in which they were grouped and treated as one composite set.

40

4. Chapter 4: Hofstede’s Model of Cultural Dimensions The first section of this chapter provides an extensive preview of Hofstede’s model, which includes its background, a description of the dimensions, the critique for and against the model, and then provides a description of the treatment of Arab countries, the focus of this research, in Hofstede’s model. The second section provides a description of the relationship between culture and interface design through defining globalization, internationalization and localization; the section then moves to discuss the relationship between culture and interface usability. The final section discusses cultural influence on interface design through an extensive examination of studies that applied Hofstede’s model in their investigation, with a sub-section dedicated to studies of cultural influence on Arabic web interfaces.

4.1. The Model Geert Hofstede’s model of culture2 has had a profound influence on the development of cross-cultural studies within psychology and the social sciences, in general, since its first appearance in 1980 (Kirkman, Lowe, & Gibson, 2006). The model was based on primary data from over 116,000 attitude surveys, which were extracted from a pre-existing database of employees within IBM subsidiaries in 72 (later reduced to 40) countries, and was undertaken between 1967 and 1969 and again between 1971 and 1973. The selection of the 40 countries included in the model was based on the criterion of having at least four out of seven occupational categories in the survey represented with eight or more respondents for each country. From the statistical analysis of these surveys, Hofstede formulated four dimensions of national culture to assist in differentiating cultures, on which each 2

Sometimes referred to in the literature as a theory rather than a model (e.g. Callahan, 2007; Myers & Tan, 2002)

41

country was given a comparative score (see Table 2). These dimensions comprise of: Power Distance-the extent to which the less powerful members of organizations expect and accept that power is distributed unequally; Individualism-the extent to which individuals are integrated into groups; Masculinity/Femininity-assertiveness and competitiveness versus modesty and caring; Uncertainty Avoidance-intolerance for uncertainty and ambiguity. A fifth dimension, Long-Term Orientation-the degree of future orientation-was added later when Hofstede expanded his model to include 10 more individual countries and three regions. Hofstede added a sixth dimension, Indulgence/Restraint, to the model in 2010. For the three regions, Hofstede decided to group a number of countries into regions based on the assumption of having identical cultural traits. These regions were East Africa, West Africa, and the Arabic-speaking countries (Hofstede, 2001). These dimensions are described further here: 1. Power distance (PD) describes the perception of equality and inequality by members of different cultures, where cultures with high PD tend to have centralized political power, exhibit tall hierarchies in organizations, and have large differences in salary and status among employees. Low PD cultures tend to view subordinates and supervisors as closer together and more interchangeable, have flatter hierarchies in organizations, and less differences in salaries and status among employees. 2. Uncertainty Avoidance (UA) describes the extent to which the members of a culture feel threatened by unknown situations. Cultures with high UA tend to be expressive, and its people tend to talk with their hands, raise their voices, and show emotions. They also seem active, emotional or aggressive and avoid ambiguous situations. On the other hand, cultures with low UA tend to be less expressive, less openly anxious, and people behave quietly without showing aggression or strong emotions.

42

Table 2. Index scores and ranks for countries and regions (Hofstede, 2001, p. 500) Dimension 1 Dimension 2 Dimension 3 Power Uncertainty Individualism Country Distance Avoidance Index Rank Index Rank Index Rank Argentina 49 35-36 86 10-15 46 22-23 Australia 36 41 51 37 90 2 Austria 11 53 70 24-25 55 18 Belgium 65 20 94 5-6 75 8 Brazil 69 14 76 21-22 38 26-27 Canada 39 39 48 41-42 80 4-5 Chile 63 24-25 86 10-15 23 38 Colombia 67 17 80 20 13 49 Costa Rica 35 42-44 86 10-15 15 46 Denmark 18 51 23 51 74 9 Ecuador 78 8-9 67 28 8 52 Finland 33 46 59 31-32 63 17 France 68 15-16 86 10-15 71 10-11 Germany 35 42-44 65 29 67 15 Britain 35 42-44 35 47-48 89 3 Greece 60 27-28 112 1 35 30 Guatemala 95 2-3 101 3 6 53 Hong Kong 68 15-16 29 49-50 25 37 Indonesia 78 8-9 48 41-42 14 47-48 India 77 10-11 40 45 48 21 Iran 58 29-30 59 31-32 41 24 Ireland 28 49 35 47-48 70 12 Israel 13 52 81 19 54 19 Italy 50 34 75 23 76 7 Jamaica 45 37 13 52 39 25 Japan 54 33 92 7 46 22-23 S. Korea 60 27-28 85 16-17 18 43 Malaysia 104 1 36 46 26 36 Mexico 81 5-6 82 18 30 32 Netherlands 38 40 53 35 80 4-5 Norway 31 47-48 50 38 69 13 New Zealand 22 50 49 39-40 79 6 Pakistan 55 32 70 24-25 14 47-48 Panama 95 2-3 86 10-15 11 51 Peru 64 21-23 87 9 16 45 Philippines 94 4 44 44 32 31 Portugal 63 24-25 104 2 27 33-35 S. Africa 49 35-36 49 39-40 65 16 Salvador 66 18-19 94 5-6 19 42 Singapore 74 13 8 53 20 39-41 Spain 57 31 86 10-15 51 20 Sweden 31 47-48 29 49-50 71 10-11 Switzerland 34 45 58 33 68 14 Taiwan 58 29-30 69 26 17 44 Thailand 64 21-23 64 30 20 39-41 Turkey 66 18-19 85 16-17 37 28 Uruguay 61 26 100 4 36 29 United States 40 38 46 43 91 1 Venezuela 81 5-6 76 21-22 12 50 Yugoslavia 76 12 88 8 27 33-35 Regions: Arab Countries 80 7 68 27 38 27-27 E. Africa 64 21-23 52 36 27 33-35 W. Africa 77 10-11 54 34 20 39-41

43

Dimension 4 Masculinity Index 56 61 79 54 49 52 28 64 21 16 63 26 43 66 66 57 37 57 46 56 43 68 47 70 68 95 39 50 69 14 8 58 50 44 42 64 31 63 40 48 42 5 70 45 34 45 38 62 73 21

Rank 20-21 16 2 22 27 24 46 11-12 48-49 50 13-14 47 35-36 9-10 9-10 18-19 43 18-19 30-31 20-21 35-36 7-8 29 4-5 7-8 1 41 25-26 6 51 52 17 25-26 34 37-38 11-12 45 13-14 40 28 37-38 53 4-5 32-33 44 32-33 42 15 3 48-49

53 41 46

23 39 30-31

Dimension 5 Long-Term Orientation Index Rank 31 22-24 31 22-24 38 18 65 6 23 30 46 10 42 14 39 17 31 22-24 25 28-29 96 2 61 7 43 13 34 19 80 4 75 5 44 11-12 44 11-12 30 25-26 0 34 19 31-32 30 25-26 48 9 19 31-32 33 20 40 15-16 87 3 56 8 29 27 25 16

28-29 33

3. The Individualism/Collectivism (IND) dimension ranks cultures based on individual or collectivistic orientations of the members, where individualistic cultures value personal time, freedom, and challenge, and also have extrinsic motivators as material rewards at work. Conversely, collectivist cultures value training, physical conditions, and skills, and have intrinsic rewards of mastery. 4. Masculinity/Femininity (MAS) refers to gender roles, not physical characteristics, and is primarily characterized by the levels of the preference for achievement, heroism, assertiveness, and material success in masculine cultures verses the preference for relationships, modesty, and quality of life in feminine cultures. 5. Long-/Short-Term Orientation (LTO/STO) was added when Hofstede expanded his work in another key publication (Hofstede, 2001). This dimension describes future- versus historicalorientations of the culture. People in cultures with long-time orientation value the family as the prototype of all social organizations; believe that virtuous behavior to others means not treating them as one would not like to be treated; and also believe that virtuous behavior in work means trying to acquire skills and education, working hard, and being frugal, patient, and persevering. People in cultures with short-time orientation promote equal relationships; emphasize individualism and focus on treating others as you would like to be treated; they also find fulfillment through creativity and self-actualization. According to Fang (2003), this dimension in particular is relatively difficult to understand, and that may explain why it is not heavily used in cross-cultural research. 6. The final dimension, Indulgence/Restraint, was added to the model in 2010 based on data analysis for World Values Survey conducted by Micheal Minkov (Hofstede, Hofstede, & Minkov, 2010). Indulgence stands for a society that allows relatively free gratification of

44

basic and natural human drives related to enjoying life and having fun. Restraint stands for a society that suppresses gratification of needs and regulates it by means of strict social norms. Considering that this dimension was added after this study was designed, and that it has not been discussed in cross-cultural web design yet, it will not be further discussed or included in this research.

Figure 2. Hofstede’s model of cultural dimensions

Research based on Hofstede’s model (Figure 2) was analyzed and synthesized by Søndergaard (1994) and Yates and Cutler (1996). Søndergaard’s study showed that there were four distinct ways in which Hofstede’s model was applied: nominal quotations, citations covering remarks on Hofstede’s work such as reviews and criticisms, empirical usages of the 45

model either by duplication or minimal adjustment of the framework, and using Hofstede’s concepts as a paradigm and a set of assumptions taken for granted. According to Søndergaard’s study of the Social Science Citation Index, 1036 quotations and citations were counted from Hofstede’s Culture’s Consequences in journals during the period 1980 to September 1993. Søndergaard also recorded 36 book reviews for ‘Culture’s Consequences’ alone. As for the replication of Hofstede’s model, a total of 61 replications have been recorded. These applications of Hofstede’s findings either tried to find support for the cultural differences in other populations, or to validate the dimensions by interpreting the differences according to the same cultural dimensions. In addition, 28 replications compared matched populations and involved comparisons with paired sets of at least two different populations; 16 studies researched single populations and 13 studies used different research designs and instruments. Later, Sivakumar and Nakata (2001) noted that Culture’s Consequences has been cited 1101 times from 1987 to 1997 according to the social sciences citations index; in contrast, Edward Hall’s Beyond Culture which presents a rival culture theory, has been referenced 147 times. Smith (2002) stated that very few researchers have achieved the impact Hofstede had with his IBM study through its replication and, therefore, the model should be inspected further. Yates and Cutler (1996) examined the extent to which Hofstede's model has been applied within a business context. From a total of 54 articles published in scholarly business journals within this discipline since 1980, they concluded that Hofstede's model is indeed robust and has been successfully used to conduct both empirical and conceptual research across a great breadth of subjects.

46

When Hofstede expanded his initial work in a second edition (2001), Søndergaard (2001) reviewed it and noted that this work was a necessary, timely continuation of Culture’s Consequences, and it filled the gap between 1980 and 2000, when the first and the second editions came out. The contribution, according to Søndergaard, is in the review of published and unpublished studies that relate to the original IBM study. In doing so, Hofstede answered many of the reservations and critiques regarding the validity and reliability of the research results that he presented in the first study. Even though Hofstede’s model has been extensively used in the literature, it has also received some methodological and theoretical criticisms over its basis. A central critique of Hofstede's model is that it reduces culture to an overly simplistic five-dimension conceptualization, limiting the sample to a single multinational corporation, failing to capture the flexibility of culture over time, and ignoring within-country cultural heterogeneity and the diffusion and dynamism of both national and ethnical shifts within cultures (Baskerville-Morley, 2005; Baskerville, 2003; Fang, 2003; Kirkman et al., 2006; McSweeney, 2002). Moreover, the model was formulated based on data collected in the 1960s and early 1970s, which opens the question of whether there could have been any changes to the dimensions over the years (Fernandez, Carlson, Stepina, & Nicholson, 1997; McSweeney, 2002). Komlodi (2005) also stated that Hofstede’s model has its own inherent bias as it had been developed in western cultures and exhibited the biases of these cultures, except for one of Long/Short-term Orientation which was based on Chinese culture. However, this dimension was criticized by Fang (2003) for being inaccurate in reflecting the Chinese culture. However, the toughest critiques that the model received were from McSweeney (2002) when he questioned whether Hofstede actually identified and studied national cultures by studying the subsidiaries of 47

just one company. McSweeney pointed out that even nations, let alone companies, are not the best units for studying cultures. While not discussed at length as a shortcoming in the literature, in three instances Hofstede grouped together several countries into three different groups: East Africa, West Africa, and Arabic-speaking countries. East Africa comprised of: Ethiopia, Kenya, Tanzania, and Zambia. West Africa included: Ghana, Nigeria, and Sierra Leone. Finally, Arabic-speaking countries comprised of: Egypt, Lebanon, Saudi Arabia, Libya, Kuwait, Iraq, and the United Arab Emirates. In her analysis of university websites from 40 countries, Callahan (2007) acknowledged that Hofstede’s cluster of Arabic-speaking countries into one group is somewhat superficial. However, Callahan did not discuss the other two groups included in the model (i.e. East Africa and West Africa). The validity of Hofstede’s model was defended by other researchers, as well as by Hofstede himself (Hofstede, 2002, 2003). Hofstede acknowledged that nations may not be the best units for studying cultures, but noted that they are usually the only kind of units available for comparison and having them is better than having nothing. Hofstede also defended the dimensions, saying that any additional dimensions should be both conceptually and statistically independent from the five dimensions already defined in his model, and they should be validated by significant correlations with conceptually related external measures. As to his measurement instrument, Hofstede notes that although it is true that a survey should not be the only way to study culture, it is nonetheless a viable tool not only in his research but in cross-cultural studies in general, and refuting it would mean refuting research in other disciplines like marketing, sociology, and political science, among other domains.

48

According to Hofstede (2002), the IBM set consisted of unusually well matched samples for an unusually large number of countries. Hofstede argued that what was measured were differences between national cultures, which any set of functionally equivalent samples from national populations can supply information about. Additionally, the dimensions he formulated were based on a stable data set from IBM, which made these dimensions valid against all kinds of external measurements, and recent replications has shown this validity, at least to some extent (Ford, Kotzé, & Marcus, 2005; Kirkman et al., 2006). Williamson (2002) acknowledged some of the critiques raised by McSweeny (2002) against Hofstede’s model such as the assumption of homogeneity of cultural attributes within one country, expecting values and behaviors of individuals to be determined by their cultural background, and confusing scores of cultural dimensions with cultural constructs. However, Williamson strongly defended Hofstede’s model by warning against McSeeney’s criticism of the model from a methodological point of view, stating that his arguments about reliability and validity are insufficient to refute empirical studies that confirmed Hofstede’s model. Furthermore, Williamson warned against the danger of disregarding Hofstede’s model along with other cultural models, consequently throwing away a valuable insight before more satisfactory models have been developed. Nevertheless, Williamson calls for a need to use multiple methods from several paradigms for researching national culture to avoid the drawbacks of Hofstede’s model. Through their extensive examination of empirical research that applied Hofstede’s framework, Kirkman et al. (2006) found that large-scale studies have sustained and amplified Hofstede’s conclusions rather than contradicted them. Furthermore, their review showed that researchers have used Hofstede’s framework successfully to select countries that are culturally 49

different in order to increase variance, and that most country differences predicted by Hofstede were supported. Myers and Tan (2002) examined 36 studies in the Information Systems literature, 24 of which used some or all of Hofstede’s dimensions. Myers and Tan found Hofstede’s work has had a significant influence on management studies in general, and Information Systems research in particular. Thus, according to Myers and Tan, Hofstede’s values are clearly relevant for additional cross-cultural research. This statement is shared by Ford et al. (2005), who reject the accusation that Hofstede’s model is stereotypical in cultural usability studies. Ford et al. analyzed and synthesized the literature in this domain and concluded that Hofstede’s cultural dimensions impact a system’s usability in terms of the cognitive load, user acceptance, objective usability, and the context of use. In addition, they argued that the problem of stereotyping is not caused by the model itself, but rather, by the way it has been utilized in this domain, and therefore attest that Hofstede’s dimensions are applicable for cross-cultural interface design and usability. Additionally, despite criticizing some or all of Hofstede’s dimensions, Baskerville (2003) and Fang (2003) acknowledged that these dimensions are still influential in some disciplines more than others, as their use has increased over the years. After weighing the literature, Jones (2007) indicated that a greater argument exists to support Hofstede than dispute his work, and the majority of his findings have weathered the storms of time, and will continue to guide multinational practitioners into the global future.

4.1.1. Arab Countries in Hofstede’s Model In 1982, after the first edition of Hofstede’s Culture’s Consequences had appeared, Hofstede decided to add 10 more countries and three regions to his initial selection of 40 countries. One of the three regions comprised the Arabic-speaking countries, by which Hofstede

50

meant only Egypt, Lebanon, Libya, Kuwait, Iraq, Saudi Arabia, and the Arab Emirates (Hofstede, 2001). These seven countries were surveyed both in 1969 and 1972, but when Hofstede tried to extend the country list in 1982, it turned out that IBM had not only inadvertently wiped the tape containing the raw survey data but it had also destroyed the data printouts. The only data printouts that were saved pertained to the total region, so he was forced to treat these countries as one region, whereas he may have wanted to keep at least Egypt and Lebanon separate. This, according to Hofstede (2001, p. 52), made the region culturally less homogeneous than would be desirable. Baskerville (2005) noted that cultural homogeneity would be expected to be relatively low among Arabic-speaking countries, despite Hofstede’s claim that cultural homogeneity may be low among newer nations but not among the sample of nations analyzed in his work. The total Arabic-speaking region had 141 IBM respondents, 79 in 1969 and 62 in 1972. As a group, the seven Arab countries included in Hofstede’s model scored high on the power distance (80), the uncertainty avoidance (68), and the masculinity (53) dimensions, while scoring low on the individualism (38) dimension, indicating a collectivist culture. The only dimension that does not have any scores for these countries is the long-/short-term orientation dimension. Based on Hofstede’s description of his dimensions and due to the high power distance and uncertainty avoidance for these countries, a social system that does not allow significant upward mobility of its citizens could be expected. These countries are also highly rule-oriented with laws, regulations, and controls in order to reduce the amount of uncertainty, while inequalities of power and wealth have been allowed to grow within the society. Hofstede (2009) also points out that when these two dimensions are combined, it creates a situation where leaders have the

51

ultimate power and authority, and the rules, laws and regulations developed by those in power reinforce their own leadership and control. The high score of the masculinity dimension portrays the high levels of preference for achievement, heroism, assertiveness, and material success as opposed to preference for relationships, modesty, and quality of life. Interestingly, Hofstede (2009) has attributed the slightly high score in these countries to religion rather than to a cultural paradigm, when he proclaims that women are limited in their rights. The lowest score for Arab countries on Hofstede’s indices comes from the individualism dimension. According to Hofstede (2009), this translates into a collectivist society and is manifested in a close commitment to the group where loyalty is paramount, and overrides most other societal rules. These countries value training and skills, and have intrinsic rewards of mastery.

4.2. Relationship between the Dimensions and Web Interface Design Despite the criticism that Hofstede’s dimensions received, researchers have favored them because of their clarity, resonance, and applicability in numerous research areas. Even in information technology fields, like information systems and HCI, these dimensions have been widely used as the basis for comparing behavior across cultures (Komlodi, 2005). Although Hofstede’s dimensions originally had nothing to do with interface design, some researchers have applied them to understand cultural presence on interfaces (Downey, Wentling, Wentling, & Wadsworth, 2005; Marcus & Gould, 2000; Robbins & Stylianou, 2002, 2003; Zahedi et al., 2001).

52

The ubiquity of Hofstede's model in the field of cross-cultural web design has led many authors to reference it as a theory as well as a model. For this research, it provides a conceptual framework and supplies theoretical elements related to the fields of usability. Therefore, it is referred to as both a model and a theory.

4.2.1. Power Distance Power distance according to Hofstede (1980, 2001) is the extent of equal or unequal distribution of power between members within a culture. In web design, high power distance is reflected in the strong focus on expertise and authority of leaders, professional expertise and highly respected figures in the community, either in written communication using national and religious social models or in images and pictures (Marcus & Gould, 2000; Robbins & Stylianou, 2002; Zahedi et al., 2001). Ackerman (2002) notes that when images are used as metaphors in interfaces, they tend to depict monuments and buildings more frequently than people. Moreover, the importance of security and restrictions to access information is explicit and enforced through the use of passwords (Ackerman, 2002; Marcus & Gould, 2000). Web interfaces with low power distance, on the other hand, have less emphasis on the social and moral orders and their symbols, therefore are not significant or are in minor use. The strong focus of these interfaces is on citizens or customers rather than expertise and authority as in high power distance interfaces. In addition there is a transparent and implicit freedom to roam for users.

4.2.2. Uncertainty Avoidance Hofstede (1980, 2001) described uncertainty avoidance as the extent to which the members of a culture feel threatened by unknown situations. Web interfaces with high uncertainty avoidance are described as being simple with clear metaphors, limited choices and 53

restricted amount of data. These interfaces also attempt to reveal or forecast the results of actions through navigation schemes and redundant cues (i.e. color, typography, etc.) before users act, in order to reduce ambiguity and prevent them from becoming lost in the website (Ackerman, 2002; Marcus & Gould, 2000). Conversely, interfaces with low uncertainty avoidance are described as being complex with maximal content and choices, and also have an acceptance of wandering and risk. The focus on understanding underlying concepts rather than narrow tasks can be seen when there is less control in navigation, where links may open new windows leading away from the original location. In addition, coding of color and typography are used to maximize information, therefore multiple links are used without redundant cueing (Marcus & Gould, 2000).

4.2.3. Individualism/Collectivism This dimension ranks cultures based on individual or collectivistic orientations of the members (Hofstede, 1980, 2001). In web design, interfaces with high individualism maximize users’ motivation based on personal achievement, where images of success are depicted through materialism. These interfaces also give prominence to youth and action represented by individuals who show what is new and unique about their cultures and use controversial speech in their communication style. In addition, these interfaces provide the means for users to willingly provide personal information (Marcus & Gould, 2000). Additionally, individualistic countries would have higher presence of privacy policy statements on their websites since they believe everyone has a right to protect privacy (Robbins & Stylianou, 2002). In contrast, interfaces with high collectivism (i.e. low individualism) maximize users’ motivation based on group achievement, where images of success are depicted through social and political agendas. Therefore, the prominence in these interfaces is given to groups that underlie a sense of social morality through relationships. Aged, experienced and wise leaders are also prominent in 54

interfaces with high collectivism, in which they portray tradition and history and use official slogans in their communication style. As to personal information, these interfaces have high protection of personal data for their users.

4.2.4. Masculinity/Femininity Masculinity in interfaces was described by Marcus and Gould (2000) through traditional gender, family, and age distinctions between users. The navigation within these interfaces must be oriented to exploration and control, where the attention is gained through games and competition, and where graphics, sound, and animation are used for utilitarian purposes. Therefore, the work tasks and roles are limited and adjusted for quick results. Feminine interfaces, on the other hand, have blurring gender roles, where tasks are accomplished through mutual cooperation, exchange, and support rather than mastery and winning. In addition, attention is gained through poetry and visual aesthetics, where it appeals to unifying values of the culture.

4.2.5. Long-/Short-Term Orientation This dimension describes historical versus future orientations of the culture. In web design, the content of interfaces with long-term orientation focus on patience in achieving results, and on using cultural markers such as national images, flag colors and pictures of groups inviting participation and close social distance (Ackerman 2002; Marcus & Gould, 2000). On the other hand, interfaces with short-time orientation focus on immediate results and achievement of goals for users when seeking information from these websites.

55

4.3. Cultural Influence on Interface Design The issue of the relationship between culture and web design has been discussed widely in the literature (Callahan, 2006; Cyr and Trevor-Smith, 2004; Fernandes, 1995; Marcus, 2008; Marcus & Gould, 2000; Nielsen, 2000). Although there is no agreement on whether to incorporate what we now know as objective culture, subjective culture or both into interface design (Ford & Kotzé, 2005), it is hard to contest that design reaches its full potential when designers and their users have shared culture (Gorlenko, 2006), or at the very least some cultural understanding on the designer’s part of their potential users. Therefore, regardless of the design approach adopted by the designer, the merging of culture and usability has its implications for web design. Barber and Badre (1998) called this relationship “Culturability”. In order to understand cultural presence on the web, researchers examined and identified elements used to represent it in interfaces, including national symbols, color, and spatial organization among other elements. According to Barber and Badre (1998), these elements denote a conventionalized use of the feature on websites, not an irregular feature that occurs infrequently. The design elements can also be seen in relation to Nielsen’s (2005) and Galitz’s (2002) usability guidelines mentioned earlier, where these elements should be utilized to be simple, aesthetically pleasing, adaptive to users’ needs and perspectives, comprehensible, controllable, and familiar in the concepts and language they use based on users’ existing knowledge. Three main approaches can be adopted in culturalization: globalization, internationalization and localization. These approaches are used to examine and inspect the way culture is affecting the use and design of websites, and whether a greater cultural homogeneity

56

on the web is being achieved, or, alternatively, whether different cultural groups are adopting information technology in different ways consistent with their cultures. Marcus (2008) defined Globalization as “the entire process of preparing products or services for worldwide production and consumption and includes issues at international, intercultural, and local scales. In an information-oriented society, globalization affects most computer mediated communication, which, in turn, affects [User Interface] design” (p.357). Day (1996) defined it as applying an allegedly culture-less standard to be used across different cultures, while referring to Internationalization as designing base structures for later local cultural customization. However, the terms Globalization and Internationalization in the context of interface design are often used interchangeably and sometimes this leads to confusion. This can be seen in light of Fernandes’s (1995) definition of Globalization as “the process of creating a base design that can be changed or augmented for various countries all over the world” (p.1), thus having functionality that can be easily modified and localized to create designs and features unique to a specific country or culture. This includes identifying all target cultures, designing and developing a global base that takes common design guidelines into account, overseeing the design of all localized versions of the interface, and conducting usability tests of localized versions. As this definition is used by other researchers (Del Galdo, 1990; Nielsen, 2000) to refer to the Internationalization process of interfaces, any future reference to Internationalization will use Day’s (1996) distinct definition. Localization, on the other hand, is defined as the process of adapting a product to a particular language and culture (Cyr & Trevor-Smith, 2004; Gibb & Matthaiakis, 2007). Marcus (2008) referred to it as the process of customizing the data (including language translation but potentially other changes such as graphics, colors, and icons/symbols) and resources of code 57

needed for a specific market. Although we can conclude from this definition that localization involves substituting culturally specific elements identified in the internationalization process with a local content, it actually oversees the representation of the appropriate elements of values, tastes, metaphors, speech convention, and mental models of the target culture in the interface (the subjective culture defined earlier). These elements, in contrast to easily visible ones, such as language and currency, are more difficult to identify (Callahan, 2005). Gibb and Matthaiakis (2007) see localization as a response to globalisation and the increased internationalisation of markets, where they noted that users are more likely to buy a service or product when content is in their own language. Since the goal of localizing user interfaces is to provide a technologically, linguistically and culturally neutral platform from which to launch global initiatives while allowing a framework that incorporates local content and functionality (Cyr & Trevor-Smith, 2004), the overall communication strategy should be appropriate to the audience as well (Würtz, 2005) and the website should reflect the socio-cultural, technological and economic characteristics of its relevant cultures, both in its appearance and the list of services it provides (Zahir et al., 2002). This will eventually lead to a culturally-competent multilingual website that can help corporations lower the cost of entry to international trade, increase sales, create global demands, and establish a reliable, professional and international image online (Sun, 2001). According to Sun, the localization process is carried out on two sub-levels: 1) adjusting the features of the website including translation, punctuation, dates, weights, measurements, addresses, currency and so on, to mirror the conventions and needs of the target audience on the surface level, and 2) adjusting the aesthetic appeal, images, colors, logic, functionality and communication patterns to conform to the target audience on the cultural level. It could also include using country domain 58

symbols, for example “.eg” for websites in Egypt or “.lb” for websites in Lebanon, rather than the global .com as part of the localization process, as suggested by Nielsen (2000). One aspect that should not be forgotten is that the localization of the interface repeats the creation cycle, and functions have to be evaluated by a new cultural group of users because it is considered as a new interface (Nielsen, 1990). This process will shed light on where and how users’ tasks vary in different cultures and how they are executed.

4.3.1. Studies of Cultural Influence on Interface Design Research conducted on cultural influences on user interface design have mainly focused on comparing and identifying similarities and differences between cultures; usability tests of internationalized or localized interface versions have not been given much attention as demonstrated below in an examination of studies of cultural influences on interface design. Regardless of the methodologies that have been used to evaluate cultural influences on interface design, the aim was to discover the relationship between these two concepts in order to pave the way for a better understanding of this part of human diversity on the web and to design more usable interfaces, especially for user communities other than those in North America and Western Europe (e.g. Callahan, 2007; Cyr & Trevor-Smith, 2004; Dong & Lee, 2008). Ess and Sudweeks (2005) examined several studies that used the two key cultural models of Hofstede and Hall to see to what extent these widely used frameworks for cultural analysis are useful for cross-cultural research in the computer-mediated communication environment. Ess and Sudweeks noted that despite their recognized limitations, these two models are good tools for predicting and analyzing intercultural communication online, especially for developing the graphical elements of websites. Jagne, Smith, Duncker, and Curzon (2004) categorized the

59

studies that investigated cultural models in interface design and evaluation under three main categories: 1. Theoretical studies that use existing cultural models such as the studies of Marcus and Gould (2001); 2. Empirical studies that use existing cultural models such as the study conducted by Simon (2001); and 3. Theoretical works that use existing models combined with other approaches such as the studies of Sun (2002) and Zahedi et al. (2001)

Early studies in this domain focused on investigating the relationship between culture and usability in web design, thus identifying the design elements that can be identified as culturally specific and have an impact on web usability. One of the the most important studies was that of Barber and Badre (1998), which had a pivotal point in the development of the research in this domain as it introduced the term “culturability” by merging the terms culture and usability. Barber and Badre used inspection method where they conducted systematic usability inspection of several hundred websites from different countries, languages, and genres to find if there are design elements that can be identified as culturally specific. Patterns that had emerged in their findings suggested that these design elements, later referred to as cultural markers, are culturally determined and genre specific. Although this study gives us a deeper understanding of cultural preferences on the web, they actually assumed that these preferences impact what is deemed user friendly and therefore usability issues must take on a cultural context. This statement was based on what was found in the inspection but without conducting usability studies to confirm the findings.

60

Nevertheless, this methodology of inspecting design elements has been utilized at length in the literature in different web genres (Baack & Singh, 2007; Callahan, 2006, 2007; Tang, 2011; Zahir et al., 2002; Yli-Jokipii, 2001). The focus of content analysis varied from only examining the linear texts and nonlinear hypertext genre (Yli-Jokipii, 2001), to focusing on visual design elements such as color and layout in national web portals (Zahir et al., 2002), or multiple design elements such as colors, icons, graphics, shapes, national symbols, spatial orientation, and links in other web genres (Baack & Singh, 2007; Callahan, 2006, 2007; Tang, 2011; Sun, 2001; Würtz, 2005). Other studies used more reflective methods by including users from different cultures to identify website characteristics that are culturally preferred by the actual target groups. Some of the methods used were user evaluations (Cyr & Trevor-Smith, 2004; Downey et al., 2005; Fang & Rau, 2003; Ford & Kotzé, 2005; Gevorgyan & Porter, 2008; Simon, 2001; Sun, 2001; van Greunen & Yeratziotis, 2008), eye tracking (Dong & Lee, 2008), usability testing (Masao & Kumiyo, 1996), or a combination of content analysis and user evaluation (Baack & Singh, 2007). In these studies, the choices for web genres differed greatly, as some opted for national web portals (Zahir et al., 2002), commercial portals/websites (Fang & Rau, 2003; Würtz, 2005), municipal websites (Cyr & Trevor-Smith, 2004), or university/educational websites (Callahan, 2006, 2007; Downey et al., 2005; Gevorgyan & Porter, 2008; Tang, 2011). The rationale for choosing these genres was mainly based on the assumption that they are intended for a particular culture and likely to endure and evolve over time, and thus will be responsive to and reflective of the culture and its values to some extent.

61

The number of cultures included in these studies also varied greatly, tending mostly to choose either a few number of countries to cross-analyze (Cyr & Trevor-Smith, 2004; Dong & Lee, 2008; Fang & Rau, 2003; Gevorgyan & Porter, 2008; Masao & Kumiyo, 1996; Sun, 2001; Tang, 2011; van Greunen & Yeratziotis, 2008; Yli-Jokipii, 2001), or choosing multiple cultures to be included for analysis (Baack & Singh, 2007; Barber & Badre, 1998; Callahan, 2006, 2007; Ford & Kotzé, 2005; Simon, 2001; Zahir et al., 2002; Würtz, 2005). Those studies that focused on a limited number of cultures tended to cross-analyze North American and Asian cultures in particular, as can be seen in Table 3. Studies that used content analysis in examining web interfaces revealed a diversity of design choices applied by different countries, which confirms that cultural design differences do occur on the web to different extents (Callahan, 2006, 2007; Cyr & Trevor-Smith, 2004; Tang, 2011). Results from studies that involved users in their inspection suggest that users had different perception and satisfaction levels for websites from different cultures and that they generally prefer websites designed for their own cultures (Baack & Singh, 2007; Simon, 2001). Additionally, participants employed different viewing patterns and satisfaction when viewing a web page and concluded that web design should be carried out according to the target audience’s specific cognitive style in order to enhance perception and usage of the web page (Dong & Lee, 2008; Gevorgyan & Porter, 2008). There was also evidence of gender differences within certain cultures regarding web design elements (Simon, 2001), as well as individual factors (Zahedi et al., 2001). Many researchers favored Hofstede’s model as the basis to conduct cross-cultural interface design analysis, and thus very few used other models or a combination of several models (Baack & Singh, 2007; Downey et al., 2005; Zahedi et al, 2001). Of the few studies 62

based solely on Hall’s model (presented in Chapter 2) were those of Würtz (2005) and van Greunen and Yeratziotis (2008). While the first study opted for content analysis of McDonalds websites from countries categorized as belonging to either high-culture (HC) or low-culture (LC), the second one examined the perception of 71 South African participants on their country’s egovernment website. The results from the first study indicated that websites in HC differ from websites in LC in a variety of ways (nonverbal communication, thought patterns, power distance, collectivism/individualism, time perception, and message speed). Participants in the second study showed an overall preference to low-context features in the design of websites, regardless of the genres, which in turn contradicts the classification of South African culture as lying at a middle point. There are two main shortcomings of both studies, Würtz’s being based on the qualitative subjective interpretation of a researcher from a LC and using commercial websites for evaluation, and van Greunen and Yeratziotis’s basing their study on Würtz’s work while inspecting egovernment without adapting the research design to reflect the differences between the two web genres. However, both studies shed light on differences of communication patterns between different cultures and how these differences must be taken into account when designing interfaces, including layout, use of images, photographs, and animation. Additionally, van Greunen and Yeratziotis (2008) interestingly suggest that countries in Europe and the Middle East should lean towards one side of the scale in Hall’s model and indicate that they are easier to measure due to the lack of diversity within these regions. These studies had some limitations that may have affected their results, for example having a small number of participants with limited cultural diversity (Sun, 2001) and relying on

63

one cultural model in particular (Callahan, 2006, 2007; Baack & Singh, 2007; Gevorgyan & Porter, 2008; Tang, 2011; Simon, 2001; van Greunen & Yeratziotis, 2008; Würtz, 2005). However, the findings from the above mentioned studies still provide evidence that culture plays a role in interface design on different levels. Cultural differences are manifested in different web design elements including images, colors, layout, and so on. These studies also suggest that culture is an important aspect in multilingual web pages and thus websites need to be tailored to a particular culture in order to be successful in delivering their product.

Table 3. Summary of studies of cultural influence on interface design Study Barber and Badre (1998) Zahir, et al. (2002) Sun (2001) Yli-Jokipii (2001) Callahan (2006, 2007) Tang (2011) Simon (2001) Gevorgyan and Porter (2008) Downey et al. (2005) Cyr and Trevor-Smith (2004) Ford and Kotzé (2005) Masao and Kumiyo (1996) Fang and Rau (2003) Dong and Lee (2008) Zahedi et al. (2001) Baack and Singh (2007) Würtz (2005) van Greunen and Yeratziotis (2008)

Conceptual Model Hofstede Hofstede Hofstede Hofstede Hofstede Hofstede & Nielsen Hofstede Hofstede Nisbett Hofstede & Hall Hofstede & Schwartz Hall Hall

Cultures Multi-culture Multi-culture Brazil, Germany, China Finland, England Multi-culture China Multi-culture China, USA Germany, Japan, USA Multi-culture Japan, USA China, USA China, South Korea, USA

Methodology* CA CA UE CA CA CA UE UE UE UE UE UT UE ET

Multi-culture Multi-culture South Africa

CA & UE CA UE

* CA: content analysis; UE: user evaluation; ET: eye tracking; UT: usability testing

4.3.2. Studies of Cultural Influence on Arabic Interface Design There is an increased research interest in culturally preferred design elements for Arabic web interfaces. However, a look at some of the studies that included Arab countries shows that 64

they are either following Hofstede’s principal of grouping Arab countries and are limited in the number of countries investigated, or they are focusing on investigating one Arab country, whether included in Hofstede’s model or not, with another non-Arab country. Due to the impact of Callahan’s study (2007) on this research, it is discussed in more detail at the end of this section. The main methodologies used in these studies were also either content analysis of web design elements (Al-Sedrani & Al-Khalifa, 2012; Callahan, 2007; Marcus & Hamoodi, 2009), using surveys and user evaluation of websites (Al-Badi, 2009; Aladwani, 2013; Hamoodi & Sheikh, 2008; Tolba [2010]), or mixed methods (Abdallah & Jaleel, 2013). In these studies, the choices for web genres differed greatly, as some opted for university websites (Callahan, 2007; Marcus & Hamoodi, 2009), e-commerce websites (Al-Sedrani & Al-Khalifa, 2012), egovernment portals (Aladwani, 2013), or multiple genres (Abdallah & Jaleel, 2013; Tolba, [2010]). The number of cultures included in these studies also varied greatly from three or less countries (Abdallah & Jaleel, 2013; Al-Badi, 2009; Al-Sedrani & Al-Khalifa, 2012; Hamoodi & Sheikh, 2008; Marcus & Hamoodi, 2009; Tolba [2010]), to having multiple Arab countries (Alwahaishi, Snasel, &Nehari-Talet, 2009a, 2009b), albeit not the focus of the study (Callahan, 2007; Isa, Noor, & Mehad, 2006). These studies also had some limitations that may have affected their results, for example they have a limited number of countries included for analysis/cross-analysis (Al-Sedrani & AlKhalifa, 2012; Aladwani, 2013; Hamoodi & Sheikh, 2008; Marcus & Hamoodi, 2009), they fall short of providing details about the methodology, instrument, website selection, and criteria for evaluating the websites (Hamoodi & Sheikh, 2008), unquestionably following Hofstede’s

65

principle of grouping Arab countries (Aladwani, 2013; Callahan, 2007), or simply they did not explain their results based on any cultural model (Al-Sedrani & Al-Khalifa, 2012). Additionally, Aladwani (2013) did not discuss the rationale behind choosing British users to evaluate the usefulness and quality of a Kuwaiti e-government web portal. The results of these studies, especially those using Hofstede’s model, revealed that these websites reflect Arab culture on the web, based on the characteristics that are inferred from Hofstede’s dimensions, to different extents. For example, Masculinity and Uncertainty Avoidance were strong in these websites (Al-Sedrani & Al-Khalifa, 2012; Hamoodi & Sheikh, 2008; Marcus & Hamoodi, 2009; Tolba [2010]). However, Power Distance and Collectivism varied between these studies showing contradicting results. Abdallah and Jaleel (2013) suggested that apart from the Masculinity dimension, no other cultural dimension attained a score that was in line with Hofstede’s assessment. Tolba [2010] suggested that designers do not consider cultural markers or users’ needs when designing websites for the target audience. These designers are influenced by the designs of other local and foreign websites, but also follow the directives of the organization’s board. Marcus and Hamoodi (2009) also noted the importance of providing some of these websites in English as a barrier to local Arab users, especially when no equivalent Arabic version is provided. Al-Sedrani and Al-Khalifa (2012) went on to suggest the need to omit any content or element that does not align with Islam as an official religion in Saudi Arabia such as advertising gambling on websites localized for that market. Results from Aladwani’s study (2013) showed that there were no major differences between how Kuwaiti and British users perceive the importance of the studied e-government quality attributes, but that those were significant differences between their perceptions of the performance of the studied e-government Kuwaiti 66

portal. Other studies suggested the need to consider the broader Islamic characteristics when designing websites from Islamic countries, including Arab countries (Al-Badi, 2009; Isa et al., 2006). Other considerations suggested by some researchers include Arabic language version, Islamic appropriate images, Islamic applications and content (Al-Sedrani & Al-Khalifa, 2012). For a broader perspective on the design of Arabic websites, however not from a cultural perspective, refer to Alwahaishi, Snasel, and Nehari-Talet (2009a, 2009b). Table 4 provides a summary of these studies that included Arab countries.

Table 4. Summary of studies of cultural influence on Arabic interface design Study Marcus and Hamoodi (2009) Hamoodi and Sheikh (2008) Al-Sedrani and Al-Khalifa (2012) Aladwani (2013) Abdallah and Jaleel (2013) Callahan (2007) Tolba [2010] Alwahaishi et al. (2009a, 2009b) Al-Badi (2009) Isa et al. (2006)

Conceptual Model Hofstede Hofstede Hofstede Hofstede Hofstede Hofstede Hofstede Wong

Cultures Egypt, Jordan, UAE UAE Saudi Arabia Kuwait, Britain UAE, USA Hofstede’s group Jordan Arab Gulf UK, Oman Multi-culture

Methodology* CA CA CA CA, UE CA, UE CA CA CA UE CA

* CA: content analysis; UE: user evaluation

Callahan (2007) also included some Arab countries in her study of cultural similarities and differences on university websites based on Hofstede’s model. To sample websites from Arabic-speaking countries for this study, Callahan compiled lists of universities from each country in Hofstede’s group (Egypt, Lebanon, Iraq, Kuwait, Saudi Arabia, Libya, and the UAE) into one list and selected 20 websites that were compared to other websites from 39 countries. However, the researcher did not provide details on which websites were sampled from which country, or how the results were affected by the fact that these countries were treated, yet again,

67

as one group. When it came to websites from Arabic countries, these are the results that Callahan provided in comparison with other countries: 

Arab countries had over 35% of symmetrical pages, but statistical differences between all countries, including Arab countries, could not be calculated because of values less than five in the Chi-square table



Search engines were not popular



Animated images were frequently used with 17% of the overall pictures having animation



Logos were frequently used on these websites, either on the top or the bottom of the page



Traditional logos were more frequent than modern logos



Images of people were not frequently used and rather images of buildings were more popular



For the images that depicted people, most images showed groups rather than individuals



Images of women were used far less frequently than images of men

Based on the studies mentioned earlier, those including Arab countries or not, it is concluded that the following web design elements have a cultural dimension and thus may affect the usability and performance of website interfaces: 4.3.2.1.

Language

Language is probably the first cultural element to be noticed when visiting a website. Since there are many languages around the world, it is impossible for one website to include them all, and therefore it is important to determine the target users and adapt the language

68

accordingly. There are several elements to consider when thinking about languages including scripts, size, direction/orientation, dialects, and multiple writing systems for one language. Furthermore, whatever the language, meaning should be communicated in an inoffensive manner (Fernandes, 1995). Based on an examination of websites that are created in their country’s native language, Barber and Badre (1998) found 13 different languages on 168 native-language websites, reiterating the diversity of cultures to be found on interfaces. These languages were prevalent within their particular cultural group, even though they did not elaborate on whether these websites had other foreign or multiple languages. The languages found were Afrikaans, Arabic, Catalan, Chinese, English, French, German, Hebrew, Italian, Japanese, Persian, Portuguese, and Spanish. National language was common in Japanese, German, and American websites in Cyr and Trevor-Smith’s study (2004), but it was noted that Japanese and German websites provided translation of their sites in other languages in contrast to their American equivalents, which was reported as being surprising considering that the United States has a multicultural population. This was also mentioned in Callahan’s (2006) study, who noted that American university websites provided their home pages in English only, compared to other countries examined that provided multilingual versions of their websites mainly in their native language, as well as English. Fletcher (2006) noted that in catering for the needs of different groups, simple translation from English is rarely sufficient because a culturally sensitive approach requires that the underlying concepts rather than just the words be rendered in the other language.

69

4.3.2.2.

Color

Different cultures perceive colors differently as they associate each color with various meanings. For example, Barber and Badre (1998) noted that while red is associated with happiness in China, it resembles danger in the United States and Japan, and death in Egypt. On the other hand, white represents death in China and Japan, in contrast to purity in the United States and joy in Egypt. Color, among other design elements, contributes to the design and content of a web page, which directly affects the way that a user interacts with the site. It may impact the user’s expectations about navigation, links and content, as well as overall satisfaction through symbolic meaning, which can make a huge difference to the success or failure of the design (Fernandes, 1995). For example, red, white, and blue signify patriotism in the United States. Barber and Badre’s (1998) color inspection showed that the majority of government sites examined in their study employed national colors throughout the site, with Brazil as an exception where multicolors were used throughout most of its websites regardless of the genre. Cyr and Trevor-Smith (2004) also found considerable variation related to the use of color between Japanese, American, and German websites. For example, Japanese used the color red twice as much as the Germans or Americans, although this color resembles danger or anger in Japanese culture (Barber & Badre, 1998). Alternately, some similarities prevail as most websites used white for background followed by gray. For links, the preferred colors were blue, bluish purple, and black. Callahan’s (2006) color investigation supported the findings of the two previous studies. For example, most Greek websites were dominated by two colors; white and blue representing

70

its national colors. Websites from Ecuador used brighter colors on their interfaces. And even though Würtz’s (2005) study did not include colors in its investigation, she did note that McDonalds’ Asian websites seemed to use brighter colors than their Western counterparts. 4.3.2.3.

Images and Icons

The use of icons on an interface is to communicate ideas without text; but these icons are as diverse as the countries and cultures around the world. Additionally, while some symbols work internationally, others do not. Designing an interface with mailboxes, for example, may be problematic since a mailbox has many different looks in different parts of the world. While the Japanese mailbox looks like a US trash can, a UK trash can looks like a US mailbox. By maximizing images that work similarly in all target locales it, of course, takes less time to localize an interface and the product will be similar to other localized versions all over the world (Fernandes, 1995). Barber and Badre (1998) found that flags are used as cultural icons to identify the origin of the site. Barber and Badre also argued that flags are used to denote alternative language choices which prove to be more efficient than using textual choices. However, this might be offensive to some user communities as one flag is used to represent several dialects. English is an official and/or first language in, for example, the United States, the United Kingdom, Australia, and Canada: whose flag should be chosen to represent this language? The same applies to Arabic, which is the first language in the Arab culture, yet is mostly represented by the flag of Saudi Arabia.

71

Cyr and Trevor-Smith’s (2004) results showed that culturally specific symbols were used to a greater extent in Japan and Germany than the United States; Japanese websites, for example, used Asian characters, currency, and other culturally specific icons. Furthermore, Callahan (2006) demonstrated a significant variation in image use among countries in her study. Some, like Ecuador and Malaysia, used fast-moving animated images, while others like the United States used slide shows. Additionally, the images used depicted different concepts like persons vs. buildings, or groups vs. individuals. In her study on High-Context vs. Low-Context image representations on the Web, Würtz’s (2005) findings demonstrated a huge difference between these two types of cultures when it came to using images, as HC websites were more likely to use images to convey information, in contrast to LC websites that used text. These images were used to reflect the values characteristic of each culture, such as family values and physical activities in HC, in contrast to free and personal time on their LC counterparts. 4.3.2.4.

Layout and Orientation

The design of the layout and orientation of web pages is highly influenced by and associated with the language of the page in terms of directions and font size. English and other European languages, for example, are laid out with the text flowing from left to right, while the layout for languages that flow from right to left, like Arabic and Hebrew, is completely reversed. Barber and Badre’s (1998) inspection of websites found that Middle Eastern websites in Arabic and Hebrew have a high frequency of orienting text, links, and graphics from right-to-left, as opposed to centered or left-to-right orientation. This process is essential for web usability because the right side would be the initial focus for users from the Middle East. Fletcher (2006)

72

pointed out this problem with an example of a detergent manufacturer using the same illustration as in the UK in an advertisement designed for an Arab country. The illustrations had a dirty pile of clothes on the left, a packet of the detergent on top of a washing machine in the middle, a pile of clean clothes on the right. These illustrations were interpreted as implying that the detergent made a clean pile of clothes dirty because the pattern of reading is from right to left, despite that the text was in the local language. Even with languages that flow from left-to-right only, some variations were found between American and German websites in the study conducted by Cyr and Trevor-Smith (2004), where the placement of banners, menus, and search functions differed between these websites. 4.4. Summary This review has demonstrated that as culture influences the way in which people interact in general, it also seems to relatively influence the way in which they interact with an interface. Design elements that include colors, layout, language, icons, images, and metaphors which contribute to the design and content of a web page can directly affect the way a user interacts with an interface. These factors have been argued to be culturally determined and can impact the usability of an interface. Consequently, interface designers need to consider the cultural orientation and preferred structures of the target user community during the development process, which includes the iterative steps of planning, research, analysis, design, evaluation, documentation, and training. Taking into consideration cultural preferences and their implications, which go deeper than mere appearances, may help to achieve more desirable solutions and to determine to what extent localized designs may be better than international or

73

global ones, based on user communities’ needs. Design elements associated with each of Hofstede’s dimensions are summarized in Table 5.

Table 5. Web design elements associated with Hofstede’s dimensions Dimension Power Distance

Associated Web Design Elements Presence of social models Structured design by using symmetrical pages Images of officials (vs. images of citizens) Traditional logos Images of buildings (vs. images of people)

Uncertainty Avoidance

Simple menus Number of sections Presence of news

Individualism/Collectivism

Images of individuals (vs. images of groups) Privacy policy statements Site customization Restricted access to information

Masculinity/Femininity

Images of men (vs. images of women) Animated images Visitor counter

Long-/Short-Term Orientation

Search engines Site maps Frequently asked questions National colors

74

5. Chapter 5: Methodology This chapter describes the research methodology used in this doctoral research to answer the question of whether Hofstede’s model of cultural dimensions can explain and predict Arabic web interface design uniformly across all Arab countries. First, the overall methodological approach which includes the method used to answer the research question is presented. Second, description of the study’s sample is provided with details of the country selection, domain and website selection. Third, the unit of analysis and units of observation, the data collection procedures, and the statistical analysis procedures are explained. Finally, the chapter reviews methodological considerations for ensuring high quality data collection and analysis taken in this study, including: the codebook, coder training and the pilot study.

5.1. Method This research is based on the established culturability content analysis methodology of Barber and Badre (1998), which was also used by other researchers (e.g. Callahan, 2007; Hamoodi & Sheikh, 2007; Zahir et al., 2002). Through this method, which Barber and Badre called the culturability inspection method, the researchers, thorough systematic inspection of websites from different countries, languages, and genres, sought to find design elements that could be identified as culturally specific. The inspection focused on design elements such as colors, icons, graphics, national symbols, spatial orientation, links, etc. According to Krippendroff (2004), content analysis is a valid method used to describe trends in communication context, allowing researchers to make inferences about the patterns and differences among similar components of that communication context. This technique is used for exploratory and explanatory research, but is most often used in descriptive research (Neuman, 75

2003). The “components” in this research are the web design elements, which have been argued in previous research (e.g. Dong & Lee, 2008; Fang, & Rau, 2003) to be prevalent and possibly preferred within a particular cultural group. The focus is mainly on visual elements because websites are essentially visual media. Amant (2005) argues that factors such as layout, design, and graphics often serve as either credibility markers users use to determine if a website merits consideration or as navigational items used to access information on a website. Content analysis can occur at one of two levels of analysis: manifest or latent analysis. According to Neuman (2003), manifest analysis involves analyzing what is on the surface. For example, counting words, phrases, or other features which yield reliable quantitative data that can easily be analyzed using inferential statistics, which in this research is represented by the objective culture, the focus of this work. Latent analysis, on the other hand, involves interpreting the underlying meaning of the content. It is the more difficult of the two levels of analysis because the researcher must have a clearly stated idea about what is being measured and the outcome depends on the subjective view of the researcher. In order to avoid the researcher’s bias in analyzing the latent content, it was decided to focus on manifest analysis. Culture arguably impacts a system’s usability in terms of the cognitive load, user acceptance, objective usability, and the context of use (Ford et al., 2005). Therefore, the theoretical basis for this research is rooted in Hofstede’s model of cultural dimensions (1980, 2001), considering the model’s significant influence on Information Systems research in general, and cross-cultural web design in particular (Myers & Tan, 2002).

76

5.2. Country Selection Al-Sayyid (1973) defined the Arab nation as “that human collectivity which speaks the Arabic language, inhabits a territory called the Arab lands, and has a voluntary and spontaneous feeling of belonging to that nation” (p. 49). The League of Arab States is “[a]n association of Arab states established in 1945 to promote cooperation among member nations in matters relating to economic and social development and foreign policy” (“Arab League,” 2007). The League’s website (http://www.lasportal.org) lists 22 current member countries (see page 28). Of these 22 countries, Palestine is not recognized as an official country by the United Nations and another four have multiple official languages (Comoros, Djibouti, Mauritania, and Somalia). In order to restrict the study to Arabic websites, these five countries were excluded from this research. The decision to exclude these countries was based on the language, where Arabic was not the sole official language of the countries in question. For example, Factbook (2010) stated that Comoros has three official languages: Arabic, French, and Shikomoro. Djibouti has two official languages: French and Arabic. On the other hand, Hassaniya is the official language of Mauritania, while Somali is the official language of Somalia. Palestine was excluded from the study as it was not recognized as an official country by the United Nations while preparing for this research 3. Moreover, Comoros, Djibouti, Mauritania, and Somalia have the lowest Internet penetration rates among the Arab countries and therefore have a very low presence on the Internet (see section 3.3.7. Information and Communication Technology in Arab Countries). This left 17 Arab countries chosen to be included in this research. Among these countries selected for analysis are the same seven included in Hofstede’s model, in addition to 10 other Arab countries, referred to as non-Hofstede group in this research. This number of countries is

3

Palestine was recognized as a non-member observer state by the United Nations on 29 November 2012

77

large enough to allow thorough investigation and yield meaningful results. Also, it is more than Hofstede initially included in his model. The final selection comprised: Algeria, Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Saudi Arabia, Sudan4, Syria, Tunisia, Yemen, and the United Arab Emirates (UAE). Later during the data collection phase and due to the political events in the Arab countries dubbed “The Arab Spring”, the majority of websites in Libya could not be accessed due to Internet shut downs. It was eventually dropped from this research leaving a total of 16 Arab countries to be included.

5.3. Genre and Website Selection Barber and Badre (1998) defined a genre as a knowledge domain that conveys a specific type of information in a website and describes large categories of sites. For example, News and Media is a knowledge domain, which presents similar types of information, but may vary stylistically, such as a magazine, newspaper, or broadcast. Typically, websites from different genres contain different content and visual elements, therefore conducting the study on a variety of web genres may introduce many variables and prevent the achievement of meaningful results (Callahan, 2007). The selection of the website domain to be included in this research needed to be based on the reflection of culture and cultural values on the web pages of organizations or institutions pertaining to that domain. Therefore, the ideal first choice was to include websites from cultural institutions such as museums or national libraries in the Arab countries, but the low number of 4

At the time of conducting this research, The Sudan was a single country, referred to as the Republic of the Sudan, which in July 2011 split into Sudan and The Republic of Southern Sudan

78

websites from some countries was the biggest challenge for this domain. For example, Jordan has 15 museums (Department of Antiques of Jordan), but none of these museums has its own website, and neither do all five museums found in Kuwait (National Council for Culture, Arts, and Literature of Kuwait). On the other hand, all four museums in Qatar have their own websites (Qatar Museums Authority). To build a representative sample across the 16 Arab countries, a decision was made to include other domains, and thus government and university genres were chosen. This study focused on government websites (Zahir et al., 2003) for two reasons. First, they are primarily intended for a particular culture or nation, rather than the worldwide Internet community; governments have committed to creating a one-stop shop of websites for their citizens (Padilla, 2008). These websites are defined as web “supersites” that provide a variety of services and aggregated information from various sources including the web itself, news, discussion groups, and links to other sites (PC Magazine, 2010). Deloitte Research (2000) defined eGovernment as “the use of technology to enhance the access to and delivery of government services to benefit citizens, business partners and employees” (p.1). Governments are rethinking their interaction with their citizens as they begin to realize that an enterprise approach to customer service yields the best results. Thus, governments use their web portals as a means to provide their citizens with customized access and services to public information. Second, indigenous portals catering to the needs of distinct cultural groups can be expected to have as one design goal the reflection of the socio-cultural, technological and economic characteristics of their intended cultures in order to be successful in the services they provide (Zahir et al., 2002). Aladwani (2013), whose study focused on egovernment websites, also notes that the majority of past research in cross-cultural website 79

analysis focused mainly on commercial and e-business interfaces, as can be seen in the works of Singh and Baack (2004), Würtz (2005), Yli-Jokipii (2001), Xu and Fang (2011). Therefore, there is a reasonable need to explore cross-cultural similarities and differences in an egovernment website context. Since a list of government web portals does not exist for Arab countries, it was important to conduct an online search to locate these portals in order to investigate their content. On Google, the terms “government” and “portal” (in Arabic) associated with each of these countries were used to locate and identify sample Arabic government web portals. These search terms were sufficient to find all related sites pertaining to the Arab countries included in the research, except for the ones from Sudan and the UAE as they were still under development. However, the problem of not having a large enough sample size for the analysis was encountered yet again, and therefore the retrieved 15 websites were used in two ways: first, in a pilot study (Khashman & Large, 2010) to test the research methodology, and secondly to locate government websites that served as a research sample frame, assuming that other government websites would also likely reflect the socio-cultural, technological and economic characteristics of their intended cultures. National portals will reflect the culture and values of developers, but they are also ventures that must be responsive to the culture of their intended users to be successful (Zahir et al., 2002). The sample frame selected for the government genre is based on the lists of government agencies’ websites provided on the web portals of the countries’ governing body (see Table 6), since many governments have already committed to creating one-stop shop of websites for their citizens. As Padilla (2008) noted, being able to access links to different government agencies on a government's main portal results in ease of use in navigation, especially for governments with 80

many territories. It is also useful for the researcher who may not be familiar with a particular country's system of government. In the cases where websites were not provided with links or if their web address was not functional, a Google search was conducted using the country’s name associated with names of ministries to locate them on the web. The final sample frame was entered into an Excel sheet to perform the random sampling. Table 6. Government web portals Country Portal URL Egypt www.egypt.gov.eg Iraq www.cabinet.iq/ Jordan www.jordan.gov.jo Lebanon www.informs.gov.lb Saudi Arabia www.saudi.gov.sa Syria www.e.sy Yemen www.yemen.gov.ye Sudan www.sudan.gov.sd Morocco www.maroc.ma Tunisia www.ministeres.tn Kuwait www.e.gov.kw Algeria www.cg.gov.dz Bahrain www.bahrain.bh Qatar www.gov.qa Oman www.oman.om United Arab Emirates www.government.ae

For consistency, the websites of the highest governing body responsible for a particular area of activity or a known sector of government public administration, for example trade, defense, or transport, were chosen for analysis. These agencies, known as Ministries in Arab countries, can have responsibility for one or more departments, bureaus, commissions or other smaller executive or administrative organizations that provide public services. However, websites from these sub-departments were not included in this research. A simple random sampling was conducted to determine which 10 government websites for each country would be

81

included in the study. The names of ministries were entered in an Excel spreadsheet in column A. To generate the random sample, using the formula “=INDEX($A:$A,RANDBETWEEN(1,COUNTA($A:$A)),1)” in the cell where the random name returned was needed. This formula picks a name at random from the list in column A. It is also dynamic in that when one removes names from the list, another name will be automatically generated. For some countries, a number of government websites included in the sample were not functional (19%), they were: Yemen and Sudan (6 websites each), Iraq and Jordan (3 each), Syria, Tunisia, and Qatar (2 each), and Algeria (1). These were removed and replaced by other websites from the sample frame. The final sample of government websites included in the main study is provided in appendix A. Despite the argument that websites from different genres contain different content and visual elements (Barber & Badre, 1998) which may introduce many variables to achieve meaningful results, conducting the study on the government genre alone would introduce other problems for the sampled countries. The number of government websites from Arab countries might prove to be limited (Abdeen, Elsehemy, Nazmy, & Yagoub, 2011; Noruzi, 2006) and therefore not enough to yield meaningful statistical results. Based on this and following Callahan’s (2006, 2007) approach, university websites from the education genre were also included in the study. The university domain was also chosen due to the role of universities as cultural institutions. As Rajakumar (2003) stated, education is socially and culturally shaped. Thus, cultural values and preferences should be reflected in the goals and objectives of the universities and echoed in the design of their web pages. Callahan (2007), points out that universities nowadays, as institutions of education and culture, seem to be well aware of the opportunities the 82

Internet brings to develop wider international relations and attract foreign students. Thus the design of the home page as a virtual gateway to the rest of the world should be given high priority. In order to reduce any ambiguity regarding the domain selection, the term “university” hereafter will include all institutions of higher education that offer a bachelor’s degree (or its equivalent) or higher, regardless of the institution’s title or whether it is public or private. This criterion was also used to check the universities included in the sample frame when no information was provided in the International Handbook of Universities, as will be discussed later. The selection of the university websites was more complex than in the case of government websites as the number of institutions involved in the investigation was larger than the government genre. In order to determine which websites to include in the study, there was a need for a comprehensive list of all universities in the 16 selected Arab countries. First, websites of the countries’ higher education governing bodies (Ministry of Higher Education/ Ministry of Higher Education and Scientific Research, etc.) were examined, as these ministries usually post a list of universities in their countries. Second, the names of these universities were verified through the International Handbook of Universities, a publication by the International Association of Universities, to identify which universities offer bachelor’s degrees (or its equivalent) or higher. The Handbook provides detailed information about universities in 184 countries, including some of the countries selected for this study. The samples of websites are not homogenous among countries due to variation in the number of universities in each country. In other words, the population of websites for each

83

country is different. Some countries have over 50 universities as in the case of Saudi Arabia or Algeria. Other countries have far fewer, as is in the case of Kuwait and Qatar. This is attributed to size of the population of each country, where the more populated countries tend to have more universities (Callahan, 2007). For some countries, a few university websites included in the sample did not work: Iraq and Yemen (2 each), Lebanon, Syria, Morocco, Oman, the UAE, and Saudi Arabia (1 each), Sudan (4). They were removed and replaced by other websites from the sample frame. A simple random sampling was conducted to determine which 10 websites to include in the study. Replicating what had been done for the government agencies, the names of universities were entered in an Excel spreadsheet in column A. To generate the random sample of universities, the formula “=INDEX($A:$A,RANDBETWEEN(1,COUNTA($A:$A)),1)” was used in the cell where the random name returned was needed. This function picks a name at random from the list in column A. It is also dynamic in that when one removes names from the list, another name will be automatically generated (see appendix B for the final sample).

5.4. Unit of Analysis The main study employed the content analysis method, described by Krippendroff (2004) as “a research technique for making replicable and valid inferences from texts (or other meaningful matter) to the contexts of their use” (p.18). The unit of analysis is the home page of each of the selected websites, because it is argued to be the most important page on any website, getting more page views than any other page (Nielsen & Tahir, 2002). The home page also provides consistency across the samples since all units are a single page (Sun, 2001). Given its attention-grabbing and organizational roles, 84

Weare and Lin (2000) also advocate the use of the home page as a unit of analysis because it is likely to contain many central elements of web design. Callahan (2007, 2005), Würtz (2005), and Marcus and Gould (2000) have used home pages of different websites as the unit of analysis. The analysis in this study focused on graphical, organizational, and navigational elements consisting of: page layout, icons, images, colors, symmetry, page length, site maps, presence of search engines, and number of links (see pages 88-91 for the complete coding scheme, adapted from Callahan, 2007). Even though the presence of multiple linguistic versions of the home page was recorded and checked, the analysis focused on websites that were created in their country’s native language, which is Arabic, even though the site may offer other language options, except when there is no Arabic version of the website. According to Barber and Badre (1998), websites in the country of origin’s native language will depict cultural markers more specific to that particular culture, whereas a site in a non-native language will be influenced by outside cultural markers. Websites in Arabic were also chosen because researchers generally overlook websites in this language due to the lack of proficiency in it. In their study on the impact of culture on the design of Arabic websites, Marcus and Hamoodi (2009) defined “Arabic” as pertaining to the language or culture. For the purpose of this research, an “Arabic website” refers to a website created in Arabic language or created in one of the Arab countries (defined earlier) included in this work.

5.5. Units of Observation Design elements found in web pages become cultural markers when they prove to be highly prevalent within a particular cultural group and less prevalent or absent in other groups (Barber & Badre, 1998). The detailed inspection of each website was based on the codebook, a dependable system for collecting data that sets the rules on how to systematically observe and 85

record the content from the sampled websites for the main study. The codebook was also used to develop the coding sheet ( Figure 3), the form on which coders noted and listed every instance of every unit of observation (coding categories) to be analyzed. The initial codebook was based on the organizational and graphical design elements investigated and inspected by Callahan (2006, 2007). The codebook was also utilized in the pilot study to test the research methodology and then adjusted and finalized for the main study. The final codebook contained 32 variables that describe the elements found in Arabic web interfaces, some of which were identified by Callahan. Of the total number of variables, 17 described the general elements found in these interfaces (e.g. entry page, symmetry, number of links, etc.), two described the type of images (collaged or animated), and 13 described the images (e.g. logo, person, building, etc.). The final codebook is provided in Tables 7, 8, and 9.

Figure 3. Code sheet for the Ministry of National Education, Morocco

86

Elements added to/changed from Callahan’s (2007) codebook are: 

Page orientation: to avoid confusion during the coding process, the page orientation was decided upon based on the presence of the scrolling bar on the page with four possibilities: vertical, horizontal, both when having both, or neither if the page did not have any. This variable therefore does not refer to text orientation on the page



Access restriction: if the web interface had a section that required user login and password



Customization: if the web interface included an option to customize the page, for example font size, background color, etc.



Visitor counter: this element was included to indicate a number on a web page that shows how many people have visited that page



Privacy policy: this element was included to indicate the presence of a statement that declares a website's policy on collecting and releasing information about a visitor



Social models: formal or informal speeches, slogans, messages which use political or religious reference



Page length: depicts how long the page is, measured by dividing the 17” screen into quarters



Images of art and nature are added to the “Other” category

Elements in Callahan’s study not used in this research due to very low inter-coder reliability: 

Images of people > poses (descriptive)



Images of people > facial expression (smile, serious, etc.) 87



Images of people > social distance (intimate, close personal, far personal, etc.)



Images of people > action (standing, teaching, reading, etc.); and



Figurative art images representing humans since they do not exist

Table 7. Category description – web page Category

Description

Possible coding

Entry page

A page that does not contain any links except the link to the home page. Also pages that contain only links to multiple linguistic versions of the home page.

Yes/No

Page orientation

The web page design is organized to fit the screen (horizontal) or allows vertical scrolling. Pages with basic information on the first page and information in the margins would still be considered horizontal.

Vertical/Horizontal/Neither/Both

*Screen length

The number of screens the page has

Numerical

Symmetry

The web page design is vertically symmetrical or not. If a vertical line is drawn through the center of the page, the organization of both sides resembles each other.

Yes/No

Page organization

Description of how the information on the page is organized.

Description. Ex. The page is divided into four sections: one horizontal at the top (contains the name of the organization and the logo), next two columns (one with a menu) followed by the fourth, horizontal section

Menus

Description of the menus on the page.

Simple: one level menus – all options visible on the screen. Complex: menus constructed of several levels – after user places a pointer on the option, more choices appear. Page is classified as complex, if at least one menu was complex.

88

Number of links

Total number of links (including those on the menus) is counted for each page.

Numerical

Number of pictures

Total number of pictures (photos and graphics) on the home page. Buttons or menus do not count as graphics, but photos used as links do. All animated pictures are coded, regardless of their number.

Numerical

Color scheme

The color of the background and the most prominent color on the page are noted.

Description

*Restriction to access

Any use of passwords

Yes/No - Description if applicable

*Ability to customize the site

Ability to change colors, forts, font size, etc.

Yes/No

*Visitor counter

A number on a Web page that indicates how many people have visited that page

Yes/No

Presence of search engine/keyword search

A program that searches documents for specified keywords and returns a list of the documents where the keywords were found

Yes/No

Presence of site map

A link to a list of pages of a website accessible to crawlers or users

Yes/No

*Presence of privacy policy

A statement that declares a website's policy on collecting and releasing information about a visitor

Yes/No

*Presence of FAQ/Help *Presence of social models (religion/nationalism)

Yes/No Slogans, messages, etc.

Yes/No

*Presence of social media

Yes/No

*Presence of news

Yes/No

*Contact information

Yes/No

* Added to Callahan’s (2007) coding scheme

89

Table 8. Image description Category

Description

Possible coding

Collage

A photo composed from photos or graphics in which separate borders for each of the components cannot be distinguished. Pictures composed from several photos or graphics when the borders of each photo can easily be seen would be counted as single picture.

Yes/No

Animated

All animated pictures as well as slideshows activated on download or mouse click.

Yes/No

Table 9. Individual pictures – Categories Element Logo

Person

Category All graphics that display the official seal or emblem of the institution/organization

Attributes Traditional (official seal) Modern

All photos or graphics that represent contemporary people associated with institution or country: students, faculty, staff, visitors, etc. Graphics presenting sculptures or drawings of people are categorized as art. Number

Single Couple Group

Gender

Male Female Mixed (Couple/Group) Unidentified

90

Status

Official Student/Citizen Mixed (Couple/Group) Unknown

Building

All photos and graphics presenting architecture Number

Detail (part of the building) Singular Multiple Overview

Other

Images that do not include logos, persons, or buildings

Numerical and description

5.6. Data Collection Procedures Based on the list of sampled websites provided in Appendix A and B, the coding process for the selected websites began on 6 December 2010 with the Government websites from countries included in Hofstede’s model (i.e. Egypt, Iraq, Lebanon, Saudi Arabia, Kuwait, and UAE). Each website was inspected and coded in one session to ensure consistency in the data collection, averaging between 60-90 minutes for each website depending on the size and complexity of the home page. The files generated from this inspection were saved mainly in two folders pertaining to the genres included in the analysis, government and university. For each of these folders, one sub-folder was created for every country included in the research. Three main types of files were generated and saved for each website: the Excel coding sheet (Figure 3), a Word document containing screen shots from the homepage of each website, and the HTML file of the Arabic homepage. While not included in the analysis, the HTML file of the other linguistic versions of the page and the entry page were also saved, if available, for future reference. Additionally, if 91

any website was either not functioning or inaccessible, these files would still be saved and a consequent folder would be created for the substituting website, such as the case of Iraq in Figure 4. However, the data collection process was interrupted by a complete Internet shutdown in 2011 in some countries. The Arab world has witnessed a series of political demonstrations and protests which started in Tunisia in December 2010 and rapidly spread across most countries in the Arab world. In Egypt’s case, and in response to the online protests mobilization, the government turned off the Internet and SMS services on 28 January 2011 for nearly a week (Cottle, 2011), however, this did not affect data collection from Egypt as it happened before the Internet was shut down.

Figure 4. File organizations

This was not the case in Libya where the service took much longer to be restored, however, most websites were still inaccessible. Several attempts were made to inspect and code the sampled

92

websites from this country from 10 March 2011 until 22 August 2011 with no success, and therefore a decision to drop it from the research was made.

5.7. Statistical Analysis Procedures Research data can be analyzed in multiple ways, each of which would yield legitimate answers. However, there are several factors to consider in determining which statistical analysis procedure is most appropriate to perform on the variables in order to test the research hypothesis (Field, 2009), they are: 

What are the independent and dependent variables



The scale of measurement of the variable depending on their type



The number of groups in the research, where in this study each country is a considered a group



Meeting the assumptions of the statistical tests

Based on these factors, a list of the variables, their scale of measurement, and the appropriate statistical tests were decided upon for the main study. Data analyzed using a content analysis technique is usually summarized in the form of numbers, which typically require statistical procedures that describe and summarize data and then illuminate patterns from these data (Krippendorff, 2004; Riffe, Lacy, & Fico, 2005). These numbers represent the number of times each code is represented in the data set, where the raw data in the context of this research is the characteristics of web interfaces. Neuman (2003) stated that these numbers need to be coded according to a set of rules stating that certain numbers are assigned to variable attributes which are read by computers to facilitate the statistical analysis. For this reason, the data were entered into an SPSS file where each row represented the website 93

from each country and the set of columns represented the variables which are the web design elements in this study. There are two main variable types in this research which consequently decide the statistical test to be performed based on their level of measurement: categorical, nominal, interval and ratio. The categorical variables in this research comprise entry page, page orientation, symmetry, menus, search engine, site map, and other variables that have a limited number of possible values (e.g. Yes or No, Horizontal or Vertical). The level of measurement for these variables is a nominal one, which indicates the use of frequencies and percentages to describe the data, and the use of Chi-square test to examine the statistical significance when we have two groups (Field, 2009; Howell, 2010). Yates’s Continuity Correction could be generally applied to the results of the Chi-square test when they do not yield meaningful results due to having cells with expected values of less than five (Field, 2009). However, this adjustment test could not be applied in this research since the contingency tables did not have two categorical variables both of which consist of only two categories (i.e. 2x2 tables). There were no nominal or interval variables in this research, therefore the continuous variables were only at the ratio level, which have an infinite number of values that flow along a continuum. Examples of such variables in this research include: screen length, number of links, number of pictures, and number of languages. The level of measurement for these variables is ratio, which indicates the use of frequencies, the mean, and standard deviation to describe the variables, while using inferential tests to draw conclusions about the statistical differences between the groups.

To do so, the normality of distribution had to be tested to decide whether to use a parametric test or a non-parametric test for the continuous variables. To test the normality of

94

distribution, each continuous variable was examined using both graphical and statistical methods. Graphical methods included the histogram and normality Q-Q plots and statistical methods include testing the normality of each variable using the Kolmogorov-Smirnov test (KS). If the test is non-significant (p>.05), it indicates a normally distributed variable, while a significant test (p.05

0.24 30.87 1.67 7.41

1 2 1 1

p>.05 p.05 p