Social Networks Website Accessibility for People with Visual Disability

JRL JRL of the Faculty of Commerce for July 2017,Vol. 54 No. 2 Scientific Research.

Dr. Rasha Abd El Aziz Social Networks Website Associate Professor

Mr. Ahmed Hassan Accessibility for People with Teaching Associate Visual Disability: An Experi- Business Information Systems De- partment, Arab Academy for Science mental Study Technology and Maritime Transport

إمكانية وصول األشخاص ذوي اإلعاقات البصرية لمشبكات االجتماعية: دراسة تجريبية

ملخص البحث Abstract Social networks‘ accessibility is vital for إِىا١ٔةةةلوصٌٛ ةةةٛيوإٌةةةٝوصٌتةةةيىاأوص ّا ١ةةةلو ِةةةشو full participation in the society; especially حٞٛ١وٌٍّتاسولوصٌىاٍِلوفةٟوصٌّت ّة ووٚصا ةلونإٌ ةيلو for people with disability. Accordingly, the ٌألشخاصورٞٚوصإل الل.وٚناٌ اٌٟ،وفإْوصٌغشضوِٓو٘زٖو purpose of this study is to investigate how is Facebook accessible to people with visu- صٌذسصسلوٛ٘وصٌ حم١كوفٟوو١ف١لوإتاحلوٛ ٚيوصألشةخاصو al disability and the impact of enhancing رٞٚوصإل الةةلوصٌييةةش٠لوٌّٛلةة وصٌف١ ةةيٛنوٚ ةةشوت ض٠ةةضو -Facebook accessibility on their usage. In or إِىا١ٔلوصٛ ٌٛيوٌٍف١ ةيٛنو ٍةٝوصسة خذصَورٞٚوصإل الةلو der to understand the research problem, an experimental method was utilised, where صٌييش٠لوٌٍّٛل . ِٓو ًوفُٙوِتىٍلوصٌيحث،وتُوصس خذصَو fifty pretest questionnaires were distributed طش٠ملوتتش٠ي١ةل،وح١ةثوتةُوتٛص٠ة وصّ ة١ٓوصسة يألو ٍةٝو over visually impaired respondents to select صٌّ ت١ي١ٓورٜٚوصإل اللوصٌييش٠لوليًوصإلص ياسو ص ١اسو -suitable candidates for the experiment. Fi fteen respondents were found to be experts صٌّششح١ٓوصٌّٕاسي١ٓوٌٍ تشنل.وٚلذوٚ ذو ْوصّ لو تشو -in using computers, Internet, social netwo ِتةاسواو٘ةُوصيةشصيوفةةٟوصسة خذصَوصٌحٛصسة١روٚصإلٔ شٔةة و rks, and assistive technologies, and have ٚصٌتةةةةةيىاأوص ّا ١ةةةةةلوٚصٌ ىٌٕٛٛ ١ةةةةةاأوصٌّ ةةةةةا ذ ،و completed the experiment on Facebook to identify accessibility problems. After that, a ٚ ٔتةةضٚصوصٌ تشنةةلوفةةٟوصٌف١ ةةيٛنوٌٍ ةةشلو ٍةةٝوِتةةاوًو post-test questionnaire was conducted to إِىا١ٔةةلوصٌٛ ةةٛي.وٚن ةةذورٌةةه،و ةةشٞوصسةة ي١اْوِةةاون ةةذو -investigate the impact of enhancing Face ص ص ياسوٌٍ حم١ةكوفةٟو ةشوت ض٠ةضوإِىا١ٔةلوصٌٛ ةٛيوإٌةٝو book accessibility on the usage of people with visual disability. ِٛلة وفاسة١يٛنو ٍةٝوصسة خذصَوصألشةخاصورٞٚوصإل الةلو صٌييش٠ل. ظٙشأؤ ائجوصٌيحثو ْون ضو١ِةضصأوِٛلة و -The research results show that some fea tures of the Facebook website are inacces- ف١ يٛنو و٠ّىٓوصٛ ٌٛيوإٙ١ٌاوٌألشخاصورٞٚوصإل اللو sible for people with visual disability, while صٌييةةةش٠ل،وفةةةٟوحةةة١ٓو ْوص١ٌّةةةضصأوصألصةةةشٜوشةةةيٗولانٍةةةلو -other features are semiaccessible. The re ٛ ٌٍٛي.وٚلذو ي وصٌيحثو ْوصألشةخاصورٞٚوصإل الةلو search concludes that people with visual disability have received insignificant atten- صٌييةةةش٠لوٌةةةُو٠حهةةةٛصونا٘ ّةةةاَووي١ةةةشوف١ّةةةاو٠ ٍةةةكونٙةةةزٖو tion regarding these problems. Finally, the صٌّتاوً.وٚ ص١شص،و٠ت١شوصٌيحثوإٌٝولائّلوِٓوصٌ ٛ ١اأو -rese-arch suggests a list of recommenda ٌٍف١ يٛنوِةٓو ةًو ْوتىةْٛو وسةشوسةٌٛٙل. ٘ةزٖوصٌٛسلةلو -tions for Facebook in order to be more ac cessible. To sum up, this paper is one of ٟ٘وٚصحذ وِٓو ذدول١ًٍوِٓوصٌذسصساأوصٌ ٟوت طشقوإٌٝو -few studies that brings the topic of web acc ِٛضةةةٛموإِىا١ٔةةةلوصٌٛ ةةةٛيوإٌةةةٝوصإلٔ شٔةةة وِةةةٓوليةةةًو essibility by disabled people into the era صٌّ ةةةةٛل١ٓوفةةةةٟوصٌ يةةةةشوصٌةةةةزٜوت يةةةةشوف١ةةةةٗوصٌتةةةةيىاأو where social networks is the way out of the digital divide. This makes it a thorough inv- ص ّا ١لوٟ٘وصٌ ي١ًوٌٍخشٚجوِٓوصٌفتةٛ وصٌشل١ّةل.وestigation that helps social networks make ٚ ٘ةةزصوِةةاو٠ت ةةًوصٌيحةةثوتحم١مةةاوشةةاِكو٠ ةةا ذوصٌتةةيىاأو -better informed decisions. Hence, the origi ص ّا ١لو ٍٝوصتخارولشصسصأوِ ١ٕش .وٚناٌ ةاٌٟ،وفةإْو -nality of this article lies in its focus on ac ةةاٌلو٘ةةزٖوصٌّةةاد وتىّةةٓوفةةٟوتشو١ض٘ةةاو ٍةةٝوصٌتةةيىاأو .cessible social networks

ص ّا ١لوصٌ ٟو٠ّىٓوصٛ ٌٛيوإٙ١ٌا. -Keywords Web Accessibility, Visual dis ability, Social Networks, Facebook, Assis- tive Technologies. Paper type Research paper. Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

1. Introduction There are many Internet attempts to maintain equality and improved inde- The ability to access and use the In- pendence. However, a great number of ternet is becoming increasingly im- websites still mainly target the majority portant for full participation in the soci- of users, with little concern about those ety (Lorca et al, 2016); especially for with a visual disability (Wild, 2016), people with disabilities. The term ―digi- and Egyptian people with visual disa- tal divide‖ is used to describe the gap bility in particular have received negli- between those who do and do not have gible attention regarding this matter. access to information and communica- This creates a desperate need to address tions technology. Social networks web- this issue in Egypt. Therefore, the main sites play an important role in many aim of the research is to identify Face- people's lives; as they use these social book accessibility barriers that face networks to communicate with each Egyptian people with visual disability other, and to share and exchange infor- while surfing the website, and examine mation. There is a variety of popular the impact of enhancing Facebook ac- social networks; some of which are: cessibility on the usage of people with

Facebook, Twitter and YouTube. The visual disability in the Egyptian con- role of social networks in the Arab text. The overall aim of this study can countries is not only limited in socializ- be achieved by answering the research ing and entertainment, but rather ex- questions: How is Facebook accessible tends to include a wide variety of uses to Egyptian people with visual disabil- that range from national engagement ity? And what is the impact of enhanc- and political participation to business ing Facebook accessibility on the usage entrepreneurial efforts and social ch- of Egyptian people with visual disabil- ange (Arab Social Media, 2017). ity?

According to Internet Live Stats Accordingly, the following sections (2016), there are over three billion and explore literature review on visual disa- four hundred million Internet users. bility in general and web accessibility Egypt is an important country in the for people with visual ability in particu- region, due to its strategic location, and lar. After that, previous work related to notably high population, reaching over worldwide accessibility guide lines, 87 million (according to the Egyptian assistive technologies, and social net- Central Agency for Public Mobilization works applications is illustrated. Then, and Statistics) on August 2015. In Ma- the research methodology used in the rch, 2017, Facebook reported that it has current study will be demonstrated with around 1.94 billion monthly active us- elaboration on population and sampl- ers, with 156 million users in Arab co- ing, questionnaire design, research fra-

untries, 23% of which are in Egypt. Tw- mework, accessibility testing with ex- itter on the other hand, has around 313 periment design, procedures and partic- million monthly active users, with 11.1 ipants. The questionnaire results are th- million users in Arab countries, and 1.7 en shown, followed by the research dis- million active users in Egypt (Arab So- cussion, conclusion, recommendation cial Media, 2017). and future work consecutively.

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و around 90% living in low-income set- 2. Literature Review

tings (WHO, 2016). They can be divid- The World Health Organization (W- ed into: inability to see images clearly HO) has classified disability into three and distinctly, loss of visual field, ina- key definitions that are often used in- bility to detect small changes in bright- terchangeably but have different me- ness, color blindness, and sensitivity to anings: Impairment, Disability and Ha- light (Ahmed, et al., 2017). ndicap. ‗Impairment‘ refers to an abno- rmality in the anatomy, anatomical, ph- Due to the negligible attention paid ysiological, or psychological function. to people with visual disability, there is

On the other hand, ‗Disability‘ refers to no accurate or updated statistics of the any limitation of an individual‘s ability number of people with visual disability to perform an activity in a normal man- in Egypt. According ―Al-Ahram‖ new- ner. Finally, ‗Handicap‘ refers to a sho- spaper (2013), the number of people rtcoming for a person due to an impa- with disabilities in Egypt is around 20 irment or a disability that hinders him- million. This can be divided into 6 mil- /her from fulfilling of a normal role on lion with mobility impairment, 6 mil- an equal level with a peer group due to lion with intellectual disability, 4 mil- physical and social barriers (Wood, lion with hearing impairment and 4 mil-

1980; DPI, 1982; Lang, 2007). lion with visual disability. In Egypt there is lack of awareness about the visu- 2.1 Visual Disability ally disabled people and how to deal Disabilities are divided into Mobili- with them. Many parents of visually ty Impairment, Cognitive, Hearing, Sp- disabled people might deal with a low eech, and Visual Impairment. Visual im- vision person as a blind person due to pairment refers to an impairment in vi- lack of awareness. Not to mention the sion that, even with correction, affects lack of computer and internet related the educational performance. It ranges courses in schools that are especially from low vision impairment to total bl- tailored for people with visual disability indness (Zitkus, et al., 2016). The In- (Marshall, et al. 2009). ternational Classification of Diseases classifies four levels of visual functions 2.2 Web Accessibility for Pe- which are: normal vision, moderate visual impairment, severe visual impair- ople with Visual Disabil- ment, and blindness. People with mod- ity erate and severe visual impairment are ISO 9241-171 defines accessibility classified as low vision people. The last as the usability of a product, service, classification is blindness, which is the environment or facility by people with focus of this research together with the the widest range of capabilities. Ac- perception of people with visual disa- cording to W3C, web accessibility re- bility towards Facebook accessibility. fers to disabled people‘s ability to perceive, interpret, browse, and interact,

There are around 285 million people and contribute to the web (Freire, 2012). having a visual impairment worldwide; It does not only benefit people with they can be divided into 39 million bl- disabilities, but rather all people includ- ind and 246 low vision people, with ing old ones and those who interact

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و with digital information via different which is the main reason why this devices (Petrie and Bevan, 2009). Web method was used in the underlying accessibility evaluation is an assess- study. ment of how well the web can be used by people with disabilities. There are 2.3 Worldwide Accessibility different techniques that are used to Guidelines evaluate the website accessibility. Me- Web accessibility is an umbrella te- thods aim to explore accessibility prob- rm for the study of the adequacy of web lems; such as guideline violations, fail- technologies to users with special ne- ure modes, defects, or user performance eds. This adequacy can be viewed from and include: standards review, user test- two perspectives; first, a strict perspec- ing, subjective assessments and screen- tive, where accessibility means the abil- ing techniques (Yesilada et al., 2015). ity to access, and people with visual User testing is considered the most disability cannot understand the infor- effective method that can be used to mation in images. The second is a br- evaluate the web accessibility, but it oader perspective, where the term rep- requires access to a number of people resents how easily this type of users can with disabilities to represent different interact with a web page. This perspec- disability types (Dell et al, 2012). It tive has been the main direction taken also represents the different levels of by the Web Accessibility Initiative experience in using browsers and assis- (WAI). tive technologies (Aizpurua, 2016). Af- In response to the rapid change of ter selecting the sample that will repre- the web and due to the new technolo- sent people with disability, they are gies, W3C developed WCAG 2.0 as the required to perform given tasks while second version of the W3C Web Con- being observed and being asked to th- tent Accessibility Guidelines (Schmutz, ink aloud. The evaluator records the test 2016). It has four general principles of and takes notes during the test based on accessibility, 12 guidelines, 61 success the user comments (Pereira et al, 2015). criteria and still has the three priority Finally, these audio, video recordings levels that was implemented in WCAG and user comments are analyzed by the 1.0. A summary of WCAG 2.0 guide- evaluator and the list of problems that lines are shown below in table 1 (Cald- the user faces during the test is generat- well et al, 2008; Rømen and Svanæs, ed. Many studies suggest using this 2012). One of the major goals of WC- method to make sure that the website is AG 2.0 was to describe the require- fully accessible (DRC, 2004; Vigo et al, ments for Web content accessibility in a 2013). Despite the drawbacks of this technology neutral language, so that it method such as its high cost, the avail- could be applicable to any W3C or non- ability and mobility of people with dis- W3C technology, such as CSS, SMIL, ability, and the need to predefine test SVG, XML, PDF or Flash in ad-dition scenarios need to be well predefined, to HTML and XHTML (Rømen and yet it is considered very beneficial for Svanæs, 2012). Another goal of devel- its ability to accurately identify accessi- oping these guidelines is to make its bility problems that are usually experi- requirements accurately testable, to in- enced by real users (Dell et al, 2012),

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و sure the website conformance to the must be tested by qualified human test- WCAG 2.0. So All WCAG 2.0 success ers. Sometimes, a combination of com- criteria are testable. While some can be puter programs and qualified human tested by computer programs, others testers may be used.

Table 1. Principles and Guidelines Principle Guidelines 1.1 Text Alternatives: Provide text alternatives for any non-text content so that it can be changed into other forms people need, such as large print, braille, speech, symbols or simpler language. 1. Perceivable - Information and 1.2 Time-based Media: Provide alternatives for user interface components time-based media must be presentable to users 1.3 Adaptable: Create content that can be pre- in ways they can perceive sented in different ways (for example simpler layout) without losing information or 1.4 Distinguishstructure able: Make it easier for users to see and hear content including separating foreground from background 2.1 Keyboard Accessible: Make all functionality available from a keyboard 2.2 Enough Time: Provide users enough time to 2. Operable - User interface read and use content components and navigation 2.3 Seizures: Do not design content in a way that must be operable is known to cause seizures 2.4 Navigable: Provide ways to help users navigate, find content and determine where they 3.1 Readable:are Make text content readable and un- 3. Understandable - Information derstandable and the operation of user in- 3.2 Predictable: Make Web pages appear and op- terface must be understanda- erate in predictable ways ble 3.3 Input Assistance: Help users avoid and correct mistakes 4:.Robust - Content must be ro- 4.1 Compatible: Maximize compatibility with bust enough that it can be in- current and future user agents, including as- terpreted reliably by a wide sistive technologies. variety of user agents, including assistive technologies

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و 2.4 Assistive Technologies very low. Examples of OCR soft- People with visual disability are in ware are Kurzweil 1000 and Read need for a special hardware or software, Iris (Popescu et al 2014). assistive technologies, (Suwannawut,  Talking internet browsers are tech- 2013; Goldrick et al 2014) that support nologies designed to speak web pag- people with disabilities to carry out dai- es to people with visual disability or ly activities including using computers those who have difficulties in fast and the web that they would otherwise writing. Moreover, it can be used to find impossible or difficult to do. Ex- give audio instructions to the com- amples of these technologies for people puter. Example of this software is: with visual disability include: Dragon NaturallySpeaking software  A screen reader is a piece of soft- (Hersh and Johnson, 2008; Vatavu, ware which interrogates the contents 2017). of the screen on a computer and tu- rns the contents into a non-visual 3. Social Network Applicati- form. This can be presented in syn- ons thetic speech using speakers, or br- People use social networks to sup- aille using braille displays devices, port their existing offline connections or both (Edwards and Launikonis, with people that they already know, or 2008). create new online connections. Social  Refreshable braille displays are used networks provide a wonderful way to to provide users with refreshable gather pictures and thoughts and then Braille output of what is shown on share them with people, either privately the screen (Chen et al., 2014). They or publicly. Every social network web- consist of cells which each produce site has its own features depending on a touchable Braille symbol by using the purpose of it (Mislove, 2009; Kel- a piezoelectric drive to raise a num- sey, 2010). In 2000 social media re- ber of small pins. Displays consist of ceived a great boost with the witnessing lines of twenty, forty, or eighty ce- of many social networking sites spring- lls; Modern Braille cells have eight ing up. Those that were launched incl- dots and can therefore be used for uded Lunar Storm, six degrees, cywo- computer as well as literary Braille rld, ryze, and Wikipedia. In 2001, fot- (Vatavu, 2017). olog, sky blog and Friendster were lau-  Optical characters recognition soft- nched, and in 2003, Myspace, Linked- ware (OCR) converts any scanned In, lastFM, tribe.net, Hi5, etc. In 2004, paper using desktop scanner into popular names like Facebook Harvard, electronic document, and this con- Dogster and Mixi evolved. During 20- verted file can be printed in Braille 05, big names like Yahoo! 360, You- form (Vella et al, 2014). Unfortu- Tube, Cyword, and Black planet all nately, most OCR software do not emerged (Junco et al, 2011). support recognition of the Arabic papers, except one software produ- There are many social network ap- ced by Sakhr Company, which sup- plications. However, only the main and port the Arabic language but unfor- most popular social networks are Twit-

tunately its recognition accuracy is ter, MySpace, YouTube, LinkedIn, So-

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و undCloud, and Facebook. Twitter was stantly growing (Alexa, 2015). Linke- launched in 2006 as a mobile service; dIn has more than 6,800 full-time em- because its tweets must be not bigger ployees with offices in 30 cities around than hundred and forty characters. Ac- the world. All of the Fortune 500 busi- cording to the statistics shown on Twit- nesses are members at LinkedIn.com. ter‘s website, it has 288 million mo- The site is available in 200 countries in nthly active users, and 500 million Tw- 24 languages, and has around 48 mil- eets are sent daily with more than 33 lion monthly visits globally (LinkedIn, languages (Twitter, 2015). On the other 2016). Furthermore, SoundCloud is a hand, Myspace found in 2003. To show social sound platform that was launched artists and their work, the site gives pe- in 2008 allowing users to audio connect ople access to 53 million tracks and with each other and share their sounds videos. In 2006, Myspace became the to Twitter, Tumblr, Facebook and Four- most popular social networking website square. SoundCloud can be accessed in the United States, but was overtaken anywhere using the official iPhone and in 2008 by its competitor Facebook that Android apps. By January 2012, Soun- internationally became the most popu- dCloud.com reached 10 million regis- lar social networking site worldwide tered users with an increase of over (Hamdi and Abd El Aziz, 2016). Ap- seven million users since January 2011 proximately 43.2 million users visit (SoundCloud, 2016). Myspace on a monthly basis. Myspace also has a special profile for musical Facebook is a social networking artists were they can download their website launched in February 2004. entire music into mp3 songs (Natta, According to the website statistics, it 2010; Myspace, 2016). has 1.28 billion daily active users on average for March 2017, and 1.94 bil- YouTube was found in November lion monthly active users as of March, 2005, is considered the world's most 2017. Furthermore, it has 1.03 billion popular online video community. In mobile daily active users on average for November 2006, the website was ac- July 2016, and 1.13 billion mobile mo- quired by Google. In 2015, it was rec- nthly active users as of July, 2016. The- orded to have more than one billion se statistics show that approximately users. Also 300 hours of video are up- 83.5% of daily active users are outside loaded to YouTube every minute. Mo- the US and Canada (Facebook, 2017). reover, YouTube is localized in 75 co- Facebook is used to connect friends untries and available in 61 languages. through exchanging of messages, shar- Half of YouTube views are on mobile ing links and videos, uploading and devices. Over 6 billion hours of video tagging of photos and sharing statuses are watched monthly on YouTube (Y- updates. The basic features of Facebook ouTube, 2016). LinkedIn is the world‘s is the user‘s homepage; where he can largest professional network that was monitor the ―news feed‖ of his friends found in 2003. According to Alexa or his liked ―pages‖ and ―gro-ups‖, and rankings, it occupies the sixth rank in the user‘s profile page; where he has his the United States with more than 347 own ―Timeline‖, can post his status million registered members from aro- updates and actions and where he can und the globe and their number is con- also interact with his friends through

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و public messages. Facebook is cons- detect the accessibility problems of the idered the second top ranked website ―Facebook‖ and participants‘ percep- after the Google website. And it is also tion towards these problems. the first ranked website in Egypt, according to the Alexa traffic rankings Questionnaires are relatively quick (Alexa, 2017). and easy to create, code and interpret, and it allows reaching a wide range of The total number of Facebook users respondents quickly, easily and effec- in the Arab world at the beginning of tively. It is also easy to standardize. March 2017 was 156 million. Also by Think aloud (verbal) protocol, is used March 2017 the country average for to get access to what is going on in par- Facebook penetration in the Arab re- ticipants‘ heads while performing as- gion was over 39% (Arab Social Media signed test tasks (Cotton and Gresty, Report, 2017). Internet users in Egypt 2006). This method can be used during have reached 29 million users at the or after the test. During the test, partici- beginning of March 2017. Egypt con- pants reveal important hints about how tinues to constitute about a quarter of they think about the system (Rubin and all Facebook users in the Arab region Chisnell, 2008). It enables understand- (23%) and has gained the highest num- ing why a problem exists and how ber of new Facebook users since Janu- someone tries to work around it. An- ary 2014, with an increase of over 2.6 other think aloud method is done after million users in that time period. Ac- the test session ends by replaying the cording to a report by the Digital Mar- recorded test session and ask respond- keting Academy, Egypt is the second ents to explain their problems Arab country in terms of activity on social networks (Egyptian Ministry of Therefore, before undertaking the Communications and Information test, the evaluator asked participants to Technology, 2017). say anything that crosses their mind during the test session, and explain any 4. Research Methodology problem that they face when performing a specific task. The test evaluator The survey research is concerned with sampling, questionnaire design, can pause the test session and discuss this problem with the participant to and data analysis (Abd El Aziz, 2012). The purpose of an experiment is to avoid calculating the time spent in the study causal links; whether a change in discussion as part of the task completion time. After that, the evaluator takes one independent variable produces a change in another dependent variable notes of this comments and problems for each task to analyze it. In an exper- (Creswell, 2012, Saunders et al, 2009). This research seeks to investigate the iment the investigator measures the ef- accessibility problems that face people fects of an experiment which he conducts intentionally. with visual disabilities when using the ―Facebook‖. In the first stage it uses the Several studies stated that experi- survey strategy to collect information ments in web accessibility testing are about the participants. In the second the most successful methods to detect stage an experiment is conducted to accessibility problems, but they argue

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و that they are expensive and time con- people with visual disability, a small suming (Brajnik, et al., 2010; Brajnik pretest questionnaire is distributed to 50 and Giachin, 2014; Pereira et al, 2015). people with visual disability electroni- The user testing employs techniques to cally or as a hard copy. The question- collect data while observing representa- naire questions ask respondents to rate tive end users and giving them realistic their experience and usage of comput- tasks to do (Rubin and Chisnell, 2008). ers, screen readers, internet and ―Face- Accordingly, in the underlying study, book‖ website. After this questionnaire questionnaires were used to collect data results were statistically analysed, 15 about the participants, followed by an- candidates were selected to undertake other one to evaluate the accessibility the experiment according to their expe- of the ―Facebook‖ website. Finally, in rience and usage of the previously stat- this research the think aloud method ed technologies. was also used to let users state what they would like to say about the test ta- 4.2 Questionnaire Design sks and the ―Facebook‖ website, where In the underlying study a Pre-test the observer recorded these comments questionnaire form is designed to ev- and analysed them. aluate the user experience and select them before undertaking the accessibil- 4.1 Population and Sampling ity test, this questionnaire consist of 2 The researcher must select the ap- sections, the first section consists of 4 propriate sampling technique. In many questions that request the user to rate experiments, only a convenience sam- his experience and usage towards the ple is possible because the researcher computers, screen readers, Internet and must use naturally formed groups (e.g., ―Facebook‖, in order to identify the a classroom, an organization) or volun- user‘s experience to select the appropri- teers (Sekaran, 2010). Thus, the most ate participants who undertake the test. suitable technique for the current study The second section consists of 4 ques- is the convenience sampling, which tions to know the demographic infor- involves selecting individuals or groups mation of the user which are: (gender, that happen to be available and are will- age, education and occupation). Anoth- ing to participate at the time. er small Post-test questionnaire was distributed to the participants who took Since it is very difficult to obtain a the test to check their satisfaction with large representative user sample in ac- the test. It also inquires on the Face- cessibility researches (Jay et al, 2008), book usage, and whether it has become when evaluating website accessibility more accessible and they would use it by people with visual disability, a small straight away, within a year, after get- sample size can be used to represent the ting a positive feedback, or not use it at entire population because they have all. It also inquires on the their level of common strategies, needs and expecta- understanding of the experiment tasks, tions (Petrie and Bevan, 2009), and this in addition to the demographic infor- sample can be generalized to subgroup mation Gender, Age, Education, Occu- of the population. Accordingly, and due pation) to the limitations and availability of

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و Experimental Variables use only the keyboard, and people The test conducted on this study who use a screen reader. contains a set of tasks that are consid- 3. Keyboard Access and Visual Fo- ered as a basic tasks that every basic cus (all tasks): Many people cannot user can do it every day smoothly, the use a mouse and rely on the key- test consist of 15 tasks according to the board or assistive technologies to in- user preference he can choose between teract with the web. Keyboard focus 3 options to read these tasks: should be visible and should follow 1. Using printed braille paper a logical order through the page el- 2. Through word document file ements. Visible keyboard focus co- 3. Make the evaluator read the tasks uld be a border or highlight, as sh- one by one to the user own below, that moves as you tab through the web page. In this research seven variables are 4. Forms, labels, and errors (task: 1, selected to measure the accessibility of 2, 5, and 9): Labels, keyboard ac- the ―Facebook‖, five of them are adopt- cess, clear instructions, and effective ed from the web content accessibility error handling are important for guidelines. Additionally Effectiveness forms accessibility .Form fields and and Efficiency variables are adopted to other form controls usually have vis- test their effect on website accessibility. ible labels, such as "E-mail Addr- ess:" as the label for a text field. These variables are shown below When these labels are marked up as follows: correctly, people can interact with 1. Image text alternatives ("alt text") them using only the keyboard, using (task 6): Convey the purpose of an voice input, and using screen read- image, including pictures, illustra- ers. Also, the label itself becomes tions, charts, etc to people who can- clickable, which enables a person not see the image. (The text should who has difficulty clicking on small be functional and provide an equiva- radio buttons or checkboxes to click lent user experience, not necessarily anywhere on the label text. describe the image. (For example, 5. Multimedia (video, audio) alterna- appropriate text alternative for a tives (task 12): Follow the steps search button would be "sea- above for keyboard access to ensure rch", not "magnifying glass"). that the media player controls are 2. Headings (task 9, 15): Web pages labeled and keyboard accessible. often have sections of information 6. Effectiveness: It is defined as the separated by visual headings, for ex- accuracy and completeness with wh- ample, heading text is bigger and ich users achieve specified goals, bold (like "Headings" right above this variable is measured through this sentence :-). To make this work checking if the user completed the for everyone, the headings need to task or not and if he complete it be marked up. That way people can from the first trial or making more navigate to the headings —including than one trial to complete it. people who cannot use a mouse and

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و 7. Efficiency: It is defined as the re- complete each task, this time is sources expended in relation to the calculated from he was given the accuracy and completeness with wh- task until he complete it or saying ich users achieve those goals, this that he cannot complete the task. variable is measured through calculating the time that user spend to

Figure 1. Research Framework

4.3 Research Framework The above proposed theoretical would evaluate the accessibility. These framework shown in Figure 1 is mainly categories are: based on the web accessibility initiative a. Automatic Evaluation: such as Ach- that introduces the Easy Checks ate - ecker, WAVE2, bobby, SortSite and mpt. The framework adopts 5 variables Total Validator tools could be used of the Easy Checks list that affect the to automatically evaluate the acces- website accessibility. Literature review sibility of any web page and its con- also shows that Effectiveness and Eff i- formance to selected guidelines such ciency are two main variables that a f- as ―WCAG 1.0‖,―WCAG 2.0‖ and fect website accessibility, and are ac- ―section 508‖; (Abascal et al, 2004; cordingly adopted in the research fra- Fernandes et al, 2013; Vigo et al, mework. The framework also shows 2013). that website accessibility affects web- b. Manual Inspection: this require a site usability as reviewed in literature. human to evaluate the page accessibility. This method uses different te- 4.4 Accessibility Testing chniques such as: (conformance re- There are different methods for ev- views, screening techniques, subjec- aluating the accessibility, which can be tive assessments, barrier walkthr- divided into 2 categories based on who ough and user testing) (Raufi et al, 2015).

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و The researcher can use the automatic let them concentrate on the given tasks. tools or involve humans on the accessi- The evaluator conducts the experiment bility evaluation process or make a mix by observing one participant at a time, between both; which is called semi- then repeating the same experiment

automatic evaluation. Some of these with the rest of participants and discuss methods are adapted from the usability any problems that participants face evaluation methods while others are when performing the tasks using the specifically used to evaluate accessibil- think aloud verbal method.

ity (Brajnik et al., 2010). To evaluate any website for accessibility, different This type of connection is preferred methods could be used to do so but to people with visual disability to let most evaluators use the automatic test- them be willing to and comfortable wi- ing methods; due to time constraints th carrying out the experiment. The par- and in order to achieve reliable results. ticipants are allowed to do the task in People with visual disabilities have dif- an opened time and to pause the exper- ficulties to move to the testing place in iment and resume at any time. Before addition to the lack of the availability of they start the experiment, participants visually disabled people to test the we- are asked to customize the screen reader bsite, but user testing method is more settings according to their preference, efficient and reliable for detecting the and to use the web browser they prefer. accessibility barriers on a website because in this method the user try to ex- Hardware and software used in plore the hole website and detect the the experiment: the test is executed barriers that they might face while us- using laptop PC running Microsoft ing it (Pereira et al, 2015). For this rea- windows 8 professional operating sys- son the research at hand uses the user tem. The laptop specifications are Intel testing method to check the accessibil- core I5, processor at 2.5 gigahertz speed ity problems that face people with visu- with 4 gigabyte ram. An external key- al disability while using the ―Facebook‖ board was connected to avoid the dif- website. An experiment is conducted ference in the order of keys on some with people with visual disability to laptop brands. It was also equipped detect the accessibility problems using with headphones to allow users to listen the user testing method. to the screen reader clearly and to avoid any disturbance.

4.4.1 Experiment Design Installed software: Microsoft inter- The nature of te experiment is de- net explorer version 10.00.9200-.16384. signed based on a set of basic tasks that Firefox version 28.0 and Google ch- every participant can do frequently. rome version 25.0; Internet browsers to These tasks measure the accessibility of allow users choose their preferred br- ―Facebook‖ and its conformance to the owser and also to compare the accessi- ―WCAG 2.0‖ guidelines. The experime- bility problems that occurs on each nt is executed through observing pa- browser. ―Jaws for windows version rticipants at any place they choose tak- 16.0‖ screen reader. Finally, a screen ing into consideration conducting the recording application called ―BB Fl- experiment in a suitable environment to ashBack Pro Recorder version 4‖ was

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و installed to record the speech of the cebook‖ to go directly to the search

screen reader in addition to recording box ―alt+/‖. After reaching the se- what is displayed on the screen. arch box, the participant type the e- mail address then browse the results 4.5 Experiment Procedures using up or down arrow keys, then Every participant from the selected select the required result by pressing 15 participants is invited to do the test ―enter‖ key and add them. according to his preferred time and 4. Press on the ―messages‖ button me- place the researcher tried as soon as nu or the link called ―messages‖ on possible to make the place comfortable the site navigation par to open inbox to the participant and tried to make it as then open a received message and a controlled laboratory environment. post it to his wall either by copying Before starting the test the evaluator it or search for button that share it tells the user about the reason for this directly. test also he inform the user that no per- 5. Use his Facebook account to send sonal information are collected during external e-mail to ―msn‖ e-mail ac- this test in addition to the places of the count by pressing on ―compose a devices that are around him and he will new message‖ button in the same use to let him feel independent and ―messages‖ page and fill the form comfortable, these devices are the (lap- and press on the ―send‖ button to se- top, external keyboard, braille paper nd the e-mail. that contains the tasks list and head- 6. Create a new album and add speci- phones). The test consists of 15 tasks fied picture either by opening the given to users in an organized form, photos page or press on ―add photo with tasks done in the same pages or- /video‖ button that is available in the dered to let the participant do them home page and select ―create new smoothly, as shown below: album‖. 1. Sign up for a new ―Facebook‖ ac- 7. Change the privacy of the uploaded count by visiting the website to fill photo by pressing on the ―privacy‖ the sign up form and log-out from button menu that is located beside this account. Participants can fill the the photo to select the required pri- form by using the screen reader sho- vacy settings. rtcut to go quickly to the beginning 8. Search for a specified ―Facebook of the form, search for a heading ti- page‖ by going to the ―search‖ edit tled ―sign up‖ or scan the whole pa- box that is available in the site navi- ge to find the beginning of the form. gation as described in task number 2. Log-in again by entering the creden- 3, then write the name in the search tials of the account he just created box and browse the results then using the login form. when he found it he like it by press- 3. Add 2 friends to participant‘s ac- ing on ―like‖ button. count by going to the search box. He 9. Create a ―Facebook‖ group by going can go to the search box by browse to home page and browse for a head- the edit boxes on the page using the ing called ―groups‖ then navigate shortcut provided by screen reader down until he finds a link called ―cr- ―e‖ or the shortcut provided by ―Fa- eate group‖. He is then required to

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و click it or retrieve a list of all page edit box that the user can type the links using the shortcut provided by message and then press ―enter‖ to screen reader ―insert + f7‖ and fill send. the form to create the group, and fi- 15. Go to the Facebook homepage either nally click on ―create‖ button to ap- by using the Facebook shortcut ―alt ply creating the group. + 1‖ or browse the headings using 10. Invite friends to the created group screen reader shortcut ―h‖ to find the via the create group dialogue box or :home‖ heading and press on it then from the group page. search for ―what‘s on your mind‖ 11. Go to the timeline of his friend edit box using the previously de- ―Ahmed Hassan‖ either by search- scribed methods then type the re- ing using the search box or by br- quired text as a new status and post owsing his friend list and choose it after posting it he required to go to him and click enter on the link. Th- the heading contains his post then en, browse for edit box called ―wr- navigate using ―up arrow‖ key to ite something‖ using the screen re- find ―option for this story‖ button ader function that list all edit boxes menu to press on it and choose ―de- on the page or press the shortcut lete this post‖ option then press on ―e‖ to browse all edit boxes then the confirmation button to delete the write the post and navigate using post. the ―tab‖ key until reaching the ―post or share‖ button. Finally the evaluator saves the rec- 12. In the same page, browse the posts orded file and thanks the user for his of this friend using the ―h‖ key to collaboration and effort. After complet- navigate through headings then he ing all the required tests the evaluator found the link to a video to play it, start to enter the quantitated data into he press on it then scan the page for the excel sheets to export it to the statis- button called ―pause‖ or ―play‖ us- tical software in order to generate re- ing the ―b‖ key, this is a screen results. ader shortcut, and use the shortcuts provided by flash player to control 4.5.1 The Participants the video. In this study, 50 pretest question- 13. Press on the ―like‖ button located naires were distributed to evaluate the below the link of the video then na- users experience and their level of vigate down using the ―down ar- knowledge of using computers, the in- row‖ key to find the ―write a com- ternet and specially ―Facebook‖ web- ment‖ edit box to write the request- site, and the frequency at which they ed comment and publish it. use these technologies. After reviewing 14. Chat with his friend ―Ahmed Has- the results, 16 visually disabled partici- san‖ by browse the page for a but- pants were selected to undertake the ton called ―chat‖ and press on it experiment according to their ratings in then navigate using the arrow keys the previously stated technologies. This to find the required friend name and considerably small sample of partici- press enter then the focus of the pants were tested due to limitations screen reader cursor changed to the such as: time consumed to conduct the

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و experiments, availability of the users, their usage of the Internet and their us- and the fact that most people with visu- age of ―Facebook‖ social network in al disability in Alexandria, Egypt do not order to help the researcher select the actually use a desktop interface for ―Fa- appropriate participants before they cebook‖ website. undergo the experiment. The second part is distributed to participants who According to the scope of the study, undertook the experiment to know their participants selected were familiar with demographic information, their percep- computers and assistive technologies. tion towards ―Facebook‖ social net- Also, they are familiar with ―Facebook‖ work if it became more accessible, and website, its desktop interface and they their satisfaction toward the experi- use it frequently. Finally, 1 user refused ment. to complete the test, so the total participants that are included in the analysis of Table 2 below shows the percent- this research experiment are 15 partici- ages of participants‘ usage of comput- pants divided into 9 males and 6 fe- ers, screen readers, internet and ―Face- males, their age ranges from 20 to 30 book‖. Four participants were excluded years old, 8 of which are in a post grad- because they never used computers. uate studies and the remaining 7 are According to the analysis 10% of re- university graduates. Finally, all partic- spondents rated their internet usage as ipants are employees either as a cus- good, and 18% rated their ―Facebook‖ tomer service representatives or teach- usage as good. This shows that some ers. respondents use the internet for opening ―Facebook‖ website only. 8% and 10% 4.6 Questionnaire Results of participants never used computers and screen readers respectively, reveal- The user questionnaires are divid- ing that 2% of respondents use an assis- ed into two parts; the first part was dis- tive technology to use computers such tributed before the experiment to ask as screen magnifier. the participants about their knowledge on computers, assistive technologies,

Table 2 Pre-test Questionnaire Results Never Statement Fair Average Good Expert Done 8% 10% 12% 30% 40% Using Computers (4 users) (5 users) (6 users) (15users) (20 users) Using screen 10% 16% 16% 20% 38% readers (5 users) (8 users) (8 users) (10 users) (19 users) Browsing the In- 14% 24% 20% 10% 32% ternet (7 users) (12 users) (10 users) (5 users) (16 users) Accessing a social 16% 22% 12% 18% 32% network (8 users) (11 users) (6 users) (9 users) (16 users)

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و Demographic results: was then converted into numbers as the software did not understand the time Gender: 58% (29 users) are females form. The coding system used for the and 42% (21 users) are males. Age: accessibility level of tasks is divided 70% (35 users) ranges from 20 to 30 into 3 levels: years old, 20% (10 users) ranges from 1. The task is accessible if the partici- 30 to 40, 8% (4 users) ranges from 40 pant completes it from the first trial to 50 years old and 2% (1 user) ranges 2. The task is semi-accessible if the from 50 to 60, 0% (0 users) for both participant completes it after doing below 20 and above 60. Education: more than 1 trial 74% (37 users) are college graduates, 3. The task is inaccessible if the partic- 26% (13 users) are enrolled in post ipant cannot complete the task or graduate studies and 0% (0 users) for complete it in a wrong way both no qualification and undergradu- ate. Participants‘ comments during the test were documented and statistically According to the main goal of this analyzed. So a mixed method approach questionnaire; which is selecting an is used to analyze the tasks, where expert user of computers, screen read- some variables are quantitative (such ers, Internet and ―Facebook‖, 32% of as: task completion time, number of the participants‘ (16 users) were rated participants who completed the task, as experts towards the previously stated and number of trials that user took to 4 technologies. These participants were complete the task), and others are quali- selected to undertake the research ex- tative (such as: participants‘ comments periment, but 1 participant was exclud- and website content). The experiment ed from the analysis as she refused to tasks measure the accessibility of the continue the experiment after perform- website and its conformance with ing the first 2 tasks, she declared she ―WCAG 2.0‖ easy checks. These 3 lev- was no longer willing to continue the els of accessibility measure the effec- experiment and cannot perform tasks. tiveness of tasks; as shown in the results: 4.6.1 Analysis of the Exper-  The most accessible task in the test iment is task number 13 with 100% completion from the first trial so the After all the tests were conducted, most easy and accessible task for the evaluator started to analyze every people with visual disability was to test first by playing the recorded video, like and comment on any post, and then extracting its related data to  86.6% of the participants have also Microsoft excel sheet using coding sys- completed tasks number 2, 11 and tem to quantify the test results. The ex- 14, which means that visually disa- cel sheet was then exported to a statisti- bled users can login to Facebook, cal software (IBM SPSS Statistics ver- write posts and chat with their fr- sion 22.0) for analysis. The evaluator iends with few problems. also recorded the time taken by each  The chatting task was reported to user to perform every task. The time cause problems when the feature

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و was used in a page that always up- default uploader of the photos which dated with new news feeds. is not accessible.  On the other hand, the most inacces-  Task number 3 was also considered as sible tasks were task number 5 and semi-accessible with 53.3% percent- number 12 with a result of 100% ei- age of completion after performing ther not completed or wrong comple- more than 1 trial and 6.7% not com- tion, which means that the send an e- pleted, this means that the users have mail to an external e-mail address fea- some difficulties when searching for a ture and controlling any video on the friend to add to their friend lists, Facebook are not accessible to people  This problem also appeared when with visual disability as it cannot be searching for a ―Facebook page‖ on controlled through the keyboard task number 8 with 60% completion  A percentage of 93.3% of the partici- after more than 1 trial and 20% not pants could not complete task number completed the* task. 15 which means that deleting any published post on ―Facebook‖ isn‘t Table 3 describes the number of accessible through keyboard either. tasks that was completed per each of  Moreover, task number 6 was consid- the 15 participants after reviewing the ered as the most semi-accessible task efficiency of each task. The table shows with a percentage of 93.3% comple- that 0 participants could perform all the tion after performing more than 1 tri- test tasks correctly, only a maximum of al, which means that the task can be 12 out of 15 tasks (80%) were correctly performed but is associated with performed by participants which means many accessibility problems the users that the website is not fully accessible face in order to complete it such as the to any of the participants.

Table 3. Number of Completed Tasks by every user

User Number of completed tasks 1 10/15 2 9/15 3 11/15 4 10/15 5 10/15 6 10/15 7 12/15 8 11/15 9 9/15 10 10/15 11 11/15 12 12/15 13 11/15 14 11/15 15 10/15

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

Table 4 describes the task comple- time that an experiment took is 1 hour tion rate for the experiment. The aver- 36 minutes and 35 seconds, while the age time of an experiment is 0 hours, 45 minimum time was 0 hours, 30 minutes minutes and 18 seconds, the maximum and 9 seconds.

Table 4. Task Completion rate

The task that took the maximum nds. While task 2 had a maximum time was task 1 with 32 minutes, the time of around 1 minute and 25 se- minimum time was 5 minutes and an conds, minimum of around 0 mi- average of 13 minutes, followed by task nutes and 40 seconds, and an aver- 6 with maximum of 14 minutes and 47 age of 1 minute. seconds, minimum time of 1 minute and 32 seconds, and an average time of 6 4.6.2 The Easy Checks: minutes. Finally task 3 had a maximum The experiment tasks are analyzed time of 9 minutes, a minimum of 2 through comparing the accessibility of minutes, and an average of 4 minutes ―Facebook‖ website and its conform- and 30 seconds. ance with ―WCAG 2.0‖. Web accessi-  The most effective and easy task bility initiative checks, which are also was task 13 with a maximum time of called the easy checks, are related to 1 minute and 13 seconds, minimum web accessibility to different people of 0 minutes and 30 seconds, and an with different disabilities. The scope of average of 0 minutes and 45 seco- this research used selected checks that

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و are related to web accessibility to peo- the media player buttons to be la- ple with visual disability, therefore only beled correctly and accessible 5 selected checks were chosen from the through the keyboard. This check is easy checks. measured through task number 12.

The selected easy checks are: 4.6.3 Post-test Questionnaire Re-

1. Image text alternatives (“alt text”): sults requires any picture on a website to A post-test questionnaire is conduct- have a descriptive text to describe ed after completing every experiment to the picture to people with visual dis- ask if the participant intends to use Fa- ability. This check is measured th- cebook website in case it becomes more rough task 6. accessible, examine the level of the par- 2. Headings: requires the titles of any ticipants‘ understanding to the experi- section on a website to be marked up ment tasks, and know their demograph- with a virtual heading to allow peo- ic information, as shown in Table 5. ple with visual disability who uses Based on the answers of the two ques- screen readers to navigate through tions, the results of first question re- these sections in a faster way. This flects how people with visual disabili- check is measured in the experiment ties are waiting for accessibility impr- through tasks 9 and 15. ovements to allow them use the website 3. Keyboard access and visual focus: more efficiently, as of around 73% of people with visual disability cannot the users answered by ―I would use it use the mouse to deal with the com- straight away‖ if Facebook was more puter so the primary device that the accessible. The second question results people with visual disability rely on reflected that the experiment tasks was for input is the keyboard. Therefore very clear to participants since the eval- this check requires that all user ac- uator conducted the experiment one by tions must be accessible through the one, so every participant was allowed to keyboard. This check is measured ask the evaluator for more descriptions through the performance of all ex- and explanations of any task that they periment tasks. did not understand. The demographic 4. Forms, labels and errors: this check results: participants are divided into 9 requires the forms and the labels to males and 6 females, all of which their be accessible, which means the age ranges from 20 to 30 years old, 8 screen readers can read aloud the la- are enrolled in post graduate studies bels of any text box or radio button and the remaining 7 are university gr- correctly, simultaneously informing aduates. Finally, all participants are ei- the user when making an error on ther as teachers or customer service the form. This check is measured representatives. through tasks number 1, 2, 5 and 9.

5. Multimedia (video, audio) alternatives and keyboard access: requires

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و Table 5. Post-test Questionnaire results 1. If Facebook was more accessible for the visually impaired I would use it straight away 73.33% (11 users) I intend to use it within a year 20%(3 users) I intend to use it but after I get a positive feedback 6.66% (1 user) from users I will not use it at all 0% (0 users) Please rate your level of understanding the ex- 2. periment tasks Extremely clear 80% (12 users) Quite clear 20% (3 users) Moderately clear 0% (0 users) Slightly clear 0% (0 users) Not at all clear 0% (0 users) audio captcha but it was not clear. 5. Discussion Analysis, this task is considered the Task 1:‘Sign up to a new Facebook maximum task that takes time from us- account‘ is related to the check for ―Fo- er to perform because it has many steps. rms, labels, and errors‖. While trying to complete the sign up process, so-me Task 2:‗Log-in to Facebook account‘ the problem was the labeling of participants entered invalid data and did the ―e-mail‖ edit box; most users listen not know what went wrong. This is to the screen reader saying ―password‖ mainly because the only indication that the user has entered invalid data is and the cursor of the screen reader is shown by marking the invalid/missing focusing on the ―e-mail or phone‖ edit box. Furthermore, 12 participants re- field with a different color that the sc- ported that they have problems logging reen reader cannot detect. Another prob- out from their accounts; as the logout lem was that ―Facebook‖ website does not provide a short description about button which is located under the ac- the constraints for passwords such as count button menu cannot be reached using keyboard. To solve this problem the minimum or maximum length. they open their mobile website interface Another reported problem that the of ―Facebook‖ and visiting ―http://m.- captcha; ("Completely Automated Pub- facebook.com‖ to press the logout link. lic Turing test to tell Computers and Task 3: ‗Add friends to the Face- Humans Apart‖), appeared to 4 users and do not appear to the rest of the par- book account‘. The problem arises when the participant search for the e- ticipants, the screen reader cannot read this captcha. To solve this problem us- mail provided on the task, the screen ers must use Firefox web browser and reader does not read the search result popup correctly. Furthermore, the par- install an extension called (webvisum), another solution is that the user re- ticipant must have full knowledge of the friends‘ e-mail not just his name in quested help from sited person to read it for him because he tried to listen to the order to find the correct friend as the

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و visually impaired user cannot recognize Task 7: ‗Adjust the privacy settings either the profile picture or the cover making your uploaded photo available photo. A total of 9 users faced difficul- only to your friends (keyboard access)‘, ties with the search results popup as it the ‗privacy‘ button menu cannot be doesn‘t categorize the result types such accessed using the keyboard when us- as: (people, page, group and place). ing ‗Microsoft Internet Explorer‘ browser, whereas 3participants completed Task 4: ‗Open a received message and this task correctly using ‗Mozilla Fire- post it on your wall‘ the selection of the fox‘ browser. received message often creates a problem as the screen reader announces the Task 8: ‗Search for the page called text and then when the user select it the ―Vodafone Egypt‖ and like the page‘, cursor select another text. Different par- When the participants type the search ticipants searched for the ―share‖ button keyword they listen to a result saying after the message to post it on their ―Vodafone Egypt‖ which is the same timeline without selection but they did keyword, but when they press on the not find it. result they get the Facebook application of Vodafone Egypt not the Facebook Task 5: ‗Send an email to ahmed. has- page as the screen reader could not an- [email protected] email address (la- nounced the result type at the search beling, keyboard access)‘, A total of 6 result popup. As a result the partici- participants did not know this function pants had to scan the whole page to re- and the remaining 9 participants could fine the results to show Facebook pages not complete the task because the only. ―buddy message‖ edit box is not labeled and cannot be reached using the Task 9: ‗Create a group called ―Face- keyboard and the ―send‖ button was not book accessibility‖ headings (labeling accessible through keyboard. and keyboard access)‘, a total of 5 Par- ticipants failed to reach the ―create Task 6:‘Create a photo album and Up- group‖ button smoothly as it was not load the photo called ―picture‖ from marked up with a heading. your desktop (keyboard access)‘, the participants created the album correctly Task 10: ‗Invite your friends to join the but when trying to upload the pictures ―Facebook Accessibility‖ group‘, the the default photo uploader of Facebook label of the ―add members‖ edit box was not accessible so they clicked on was not read by screen reader; as a re-

―try basic uploader‖ link to upload the sult, 7 participants performed many tri- photos. The task was performed after als to add members to the group. many trials by the participants as there was no note to indicate that the default Task 11: ‗Write ―Here I am‖ on the uploader was not accessible. A total of timeline of your friend ―Ahmed Has- 13 participants claimed that the descrip- san‖‘, the task was easily performed but tion of the picture was not required so with one issue that after the user en- many photos on Facebook are not ac- tered the text and pressed the enter key cessible and people with visual disabili- to publish the post, the enter key did not ties cannot identify it. publish the post instead it inserted a

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و new line. 9 participant‘s complained time yet they cannot delete it because that they face this problem because it is the ―option‖ button menu was not ac- always changing with the different de- cessible through keyboard. However, signs of ―Facebook‖. the 1 participant who used ‗Mozilla Firefox‘ browser accomplished the task. Task 12: ‗Playing stopping, forward and rewind the video (Multimedia (vid- 6. Conclusion eo, audio) alternatives)‘, All the partici- The social network websites have pants could easily play the video by become one of the most important and pressing the link and it automatically extensive media to provide access to plays, yet all of the participants cannot news, services, entertainment, and other control the video because the ―play/- different kinds of information in peo- pause, stop, forward and rewind‖ but- ple's daily lives. Social networks usage tons was not labeled and not accessible has been increased in the last 5 years through the keyboard. 3 participants worldwide, especially in the Arab coun- commented that recently they could use tries; unfortunately many websites have this feature on the ―YouTube‖ website been designed in ways that are difficult but when they tried on ―Facebook‖ they to navigate or are inaccessible to assis- did not succeeded. tive technologies such as screen read- Task 13:‘Like and comment on this ers. Hence, this means that people with video saying ―nice video‖‘, All partici- visual disability do not have equal ac- pants performed this task correctly, yet cess to these website as the people 13 participants reported that when per- without any disability. forming the task in a page with many The most effective way to know the dynamically changing news feeds, the accessibility problems that face people page always refreshes with new posts, with visual disability is to let them try sometimes they like different post as using the website and generate the the focus of the screen reader cursor is problems from their experiment. The changed without alerting the user. results of the experiment shows that some features of the ―Facebook‖ web- Task 14: ‗Chat with Ahmed Hassan‖ site are not accessible and people with saying ―hi‖ to him (Headings)‘, The visual disabilities cannot perform them chat part is not marked up with a head- independently. These features such as ing or land mark. A total of 13 partici- controlling the video content, also edit- pants had the same problem of the con- ing or removing any shared post could tinuously updating news feeds and the not be done using the screen reader. focus of the screen reader always ch- Also the results showed that people anging, Also 7 participants reported th- with visual disability faces difficulties at the emoticons that they sent or re- when using some features that forces ceived on the chats are not read by the them to do it more times and take more screen reader. time to do, or depend on sited person to do this task for them, because designers Task 15: ‗Write ―Deleted text‖ on your fail to put appropriate text tags on some timeline and delete it‘, a total of 14 par- graphics, links, forms, or tables to de- ticipants could write a post on their scribe this content to the screen reader.

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و As a result of this findings, Egyptian cially when using new technologies people find Facebook as inaccessible, that screen readers doesn‘t support. which highlights that they are discrimi-  Controlling any video on the ―Face- nated because they do not have equal book‖ is not accessible. access to the website as sited people.  Some technologies such as ―JavaS- Also most of the studies that test the cript and AJAX‖ cause accessibility website accessibility depend on the au- problems (Ramakrishnan et al., tomated testing tools, but these tools do 2009). not detect all accessibility problems in which it can be detected by humans. If websites become accessible, peo- Moreover, they stated that there are ple with disabilities can do daily things: compatibility problems of web content children can learn adults can make a with screen readers. To sum up, people living; seniors can read about their gr- with visual disability face some prob- andchildren, and so on. Using the web, lems that can be categorized into 2 people with disabilities can do more main categories: things themselves, without having to rely on others. Therefore, accessibility Category 1: User Based Prob- for people with disabilities is more im- lems portant now than ever before. As a re-  The user was not trained to use new sult of these findings, Egyptian people technologies such as ARIA technol- with visual disability want to have eq- ogy. ual access to the ―Facebook‖ website as  Users are frustrated by the continu- sited people. Also if the website be- ous change in the design of ―Face- comes more accessible, they will use it book‖ website; as they depend on constantly. Moreover, they stated that remembering the structure and order in Egypt there is no rule that enforces of the page elements when dealing. website designers or governmental Therefore, when the interface design websites to be accessible for people changes they make extra effort to with disability. identify the new design and deal with it. 7. Recommendations and Fu-  Users are not fully aware of all the ture Work shortcuts that smooth dealing with Thus, the research at hand recom- the website, these shortcuts either mends the involvement of human users provided by the screen reader or by with visual disability in the testing pro- ―Facebook‖ website. cess of any website to evaluate its accessibility, and its compatibility with Category 2: Website Problems: assistive technologies. It also suggests  Some Website features are not ac- the formulation of a governmental regu- cessible through keyboard. lation that forces any Egyptian website  Some parts of the website are not (especially governmental websites) to marked up with a heading. be accessible to people with disabilities  Website design updates that can re- by applying the Web Content Accessi- sults in accessibility problems espe- bility Guidelines, V.2.0. Facebook‖ should also provide the description field

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و of any uploaded photo or video manda- Egyptian Context, 14 March 2012, tory for the users in order to help the LAP LAMBERT Academic Publish- people with visual disability recognize ing, ISBN-10: 384840835X, ISBN- the photo or video. It is also recom- 13: 978-3848408351 mended to develop a solution to make 3. Ahmed, T., Hoyle R., Shaffer, P., the website understand the text entered Connelly, K. Crandall, D., and Ka- in the photo description to avoid user‘s padia, A. (2017). ‗Understanding the misuse. Finally, companies specialized Physical Safety, Security, and Priva- in the assistive technologies should de- cy Concerns of People with Visual crease the prices of these technologies Impairments‘, 21(3), pp. 56-63, (e the Arabic text to speech engines) to IEEE Internet Computing, IEEE. allow the people with visual disability in the developing countries to use it. 4. Aizpurua, A., Harper, S. and Vigo, M. (2016). ‗Exploring the relation- The limitations of the study can be di- ship between web accessibility and vided into two main limitations. The user experience‘, International Jour- first is related to number of users, who nal of Human-Computer Studies, 91, have undertaken the experiment; this pp. 13-23. problem happened because a large number of people with visual disability 5. Alexa (2017). [Online]. Available cannot use the ―Facebook‖ website as it at: http://www.alexa.com/topsites. has accessibility problems. Also some Accessed: 8th JUNE 2017. of them had transportation problems. 6. Arab social media report, (2017). The second limitation is related to the Online, available at: http://www.- tested website; this study tested some mbrsg.ae/asmr-2017x. Last acce- features of only one social network ssed: [8th JUNE 2017]. website, according to the given time to the study. As a result further research 7. Brajnik G. and Giachin, C (2014). can be done in the future, by involving ‗Using sketches and storyboards to more people with visual disability or assess impact of age difference in people with other disabilities to test the user experience’, International Jou- website, by including other features of rnal of Human-Computer Studies, the same website in the test, and also by 72 (6), pp. 552-566, 2014; doi: 10- testing other social networks websites .1016/j.ijhcs.2013.12.005, pp. 552– to evaluate their accessibility. 566, Elsevier. 8. Brajnik, g., Yesilada, y., and Harper, References s. (2010). ‗Testability and validity of WCAG 2.0: the expertise effect‘, In 1. Abascal, J., Arrue, M., Fajardo, I. et Proceedings of the 12th Internation- al. (2004). ‗The use of guidelines to al ACM SIGACCESS conference on automatically verify Web accessibil- Computers and accessibility (AS- ity‘, Universal Access in the Infor- SETS '10). ACM, New York, NY, mation Society, 3(1), pp.71–79, USA, 43-50. DOI: 10.1007/s10209-003-0069-3. 9. Caldwell, B., Cooper, M., Reid, L. 2. Abd El Aziz, R. (2012). ATM Us- G. and Vanderheiden, G. (2008). age: A Stakeholder Analysis the

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و 'Web Content Accessibility Guide- ey-Facts. And at: https://newsro- lines 2.0', Web Accessibility Initia- om.fb.com/Timeline.Accessed: 8th tive (WAI), World Wide Web Con- JUNE 2017. sortium (W3C) [online], available: 16. Fernandes, N., Lopes, R. andCa- http://www.w3.org/TR/WCAG20. rriço, L. (2013). ‗On Web Accessi- Accessed: 20th JANUARY 2016. bility Evaluation Environments‘, In 10. Cotton, D. and Gresty, K. (2006). Proceedings of the International ‗Reflecting on the think-aloud me- Cross-Disciplinary Conference on thod for evaluating e-learning‘, Brit- Web Accessibility (W4A '11). ish Journal of Educational Technol- ACM, New York, NY, USA, Article ogy, 37(1), pp. 45-54, DOI: 10.11- 4, 10 pages. 11/j.1467-8535.2005.00521.x, 17. Freire, A. P. (2012) ‗Disabled peo- Wiley Online Library. ple and the Web: User-based meas- 11. Creswell, J. W. (2012). Educational urement of accessibility‘. PhD the- Research: Planning, Conducting, sis, University of York. and Evaluating Quantitative and 18. Hamdi, N. and Abd El Aziz, R. Qualitative Research, University of (2016), ‗Facebook as an Interaction Nebraska–Lincoln, Pearson. ISBN- Platform in Higher Education: The 13: 978-0-13-136739-5. Case of an Egyptian Private Univer- 12. Dell, A. G., Newton, D. A., and Pet- sity‘, The Sixth International Con- roff, J. G. (2012). Assistive Technol- ference on Social Media Technolo- ogy in the Classroom: Enhancing gies, Communication, and Informat- the School Experiences of Students ics, SOTICS 2016, ISBN: 978-1- with Disabilities, 2nd ed. Upper Sad- 61208-504-3, August 21 - 25, 2016 - dle River NJ: Pearson Education Rome, Italy, pp. 4 – 10. Inc. 19. Hersh, A. M. and Johnson, M. a. 13. Edwards, J. N. and Launikonis, B. S. (2008). Assistive Technology for (2008). ‗The accessibility and inter- Visually Impaired and Blind People, connectivity of the tubular system Springer, ISBN: 9781846288678 network in toad skeletal muscle‘. (online) – 9781846288661. The Journal of Physiology, pp. 50- 20. Jay, C., Lunn, D. and Michailidou, 77-5089, US National Library of E. (2008). ‗End User Evaluations‘, Medicine, National Institutes of In: Harper, S. and Yesilada, Y. (eds.) Health, The Psychological Society. Web Accessibility, Human Comput- 14. Egyptian Ministry of Communica- er Interaction Series, Springer Lon- tions and Information Technology don.

(2017). ICT Indicators in Brief. Av- 21. Junco, R., Heibergert, G., and Lo- ailable: ken, E. (2011). ‗The effect of Twit- http://www.mcit.gov.eg/Publication/ ter on college student engagement Publication_Summary/2062//,Last and grades‘, Journal of Computer accessed: [8th JUNE 2017]. Assisted Learning, 27 (2), pp. 119- 15. Facebook (2017), [Online]. Availa- 132. ble at: https://newsroom.fb.com-/K-

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و 22. Kelsey, T. (2010). Social Network- of Visually Impaired Users‘.Proc- ing Spaces: From Facebook to Twit- edia Computer Science, 67, pp. 77- ter and Everything In Between. 86, Elsevier. P.536. 31. Petrie, H. and Bevan, N. (2009). ‗The 23. Lang, R. (2007). ‗The Development Evaluation of Accessibility, usabil- and Critique of the Social Model of ity and user experience‘, Chapter 20, Disability‘, Leonard Cheshire Disa- The Universal Access Handbook, bility and Inclusive Development pp. 1-16, CRC Press. Centre, UCL. 32. Ramakrishnan, I.V., Mahmud, J., 24. LinkedIN (2016). [Online]. Availa- Borodin, Y., Islam, M. A., Ahmed 1, ble at: http://press.linkedin.com/abo- F., (2009). ‗Bridging the Web Ac- ut. Accessed: 20th JANUARY 2016. cessibility Divide‘, Electronic Notes in Theoretical Computer Science, 25. Lorca, P., Andres, J. and Martinez, Elsevier, pp. 107-124. A. B. (2016). ‗Does Web accessibility differ among banks?‘ World 33. Raufi, B., Ferati, M., Zenuni, X., Wide Web, 19 (3), pp. 351-373, Sp- Ajdari, J. and Ismail, F. (2015). ringer. ‗Methods and Techniques of Adap- tive Web Accessibility for the Blind 26. Marshall, C. A., Kendall, E., Gover, and Visually Impaired‘. Procedia- M., Banks, M. E. &Gover, R. M. S. Social and Behavioral Sciences, (2009). Disabilities: Insights from 195, pp. 1999-2007, Elsevier. Across Fields and Around the Wo- rld, Greenwood Publishing Group. 34. Rømen, D. and Svanæs, D. (2012). ‗Validating WCAG versions 1.0 and 27. Mislove, A. E. (2009). Online Social 2.0 through usability testing with Networks: Measurement, Analysis, disabled users‘, Universal Access in and Applications to Distributed In- the Information Society, 11(4), formation Systems, PhD Thesis, pp.375–385. Rice University, Houston, Texas. 35. Saunders, M., Lewis, P. and Thorn- 28. MySpace (2016). [Online]. Availa- hill, A. (2009). Research Methods ble at: https://myspace.com/pressr- for Business Students, Pearson Edu- oom/aboutmyspace..Accessed: 20th cation Limited, ISBN: 978-0-273- JANUARY 2016. 71686-0. 29. Natta, O. V. (2010). ―MySpace CEO 36. Schmutz, S., Sonderegger, A. and Owen Van Natta Steps Down‖, Sauer, J. (2016). ‗Implementing Re- [Online] Available at: http://techc- commendations From Web Accessi- runch.com/2010/02/10/myspace- bility Guidelines‘. Human Factors, ceo-owen-van-natta-steps-down, 58, pp. 611-629. Accessed: 08th JANUARY 2016. 37. Sekaran, U. (2010). Research Meth- 30. Pereira, L. S., Ferreira, S. B. L. and ods for Business: A Skill-Building Archambault, D. (2015). ‗Prelimi- Approach, USA: John Wiley and nary Web Accessibility Evaluation Sons, Inc. Method through the Identification of Critical Items with the Participation

Dr. Rasha Abd El Aziz , Ahmed Hassan Social Networks Website Accessibility for People…

و 38. SoundCloud (2016), [Online]. Ava- able at: http://dl.acm.org/citation- ilable at: https://soundcloud.com- .cfm?id=2461121.2461124. /imprint. Accessed: 20th JANUARY 43. WHO (2016) World Health Organi- 2016. zation online, available at: http://- 39. Suwannawut, N., (2013). Accessibil- www.who.int/mediacentre/factsheet ity Evaluation of Online Learning s/fs282/en/. Management System for Persons 44. Wild, G. (2016), Social Media and with Visual Impairment, PhD, Uni- Accessibility. In: NAH, F. F.-H. & versity Graduate School (Decem- TAN, C.-H. (eds.) ‗HCI in Business, ber). Government, and Organizations: 40. Twitter (2016), [Online]. Available eCommerce and Innovation‘, Third at: https://about.twitter.com. And at: International Conference, HCIBGO https://about.twitter.com/company. 2016, Held as Part of HCI Interna- Accessed: 20th JANUARY 2016. tional 2016, Toronto, Canada, July 17-22, 2016, Proceedings, Part I. 41. Vatavu, R. (2017). ‗Visual Impair- Cham: Springer International Pub- ments and Mobile Touchscreen In- lishing. teraction: State-of-the-Art, Causes of Visual Impairment, and Design 45. Yesilada, Y., Brajnik, G., Vigo, M., Guidelines‘, and Harper, S., (2015). ‗Exploring International Journal of Human– perceptions of web accessibility: a Computer Interaction, 33 (6): Spe- survey approach’, Behavior and In- cial Issue on Mobile Human–Co- formation Technology, 34 (2), 2015; mputer Interaction, Taylor and Fran- 10.1080/0144929X.2013.848238, cis Online, pp. 486-509. pp.119-134, Taylor and Francis online. 42. Vigo, M., Brown, J. and Conway, V. (2013). ‗Benchmarking web acces- 46. YouTube (2016), [Online]. Availa- sibility evaluation tools‘, Proceed- ble at: http://www.youtube.com/yt- ings of the 10th International Cross- /about/ and http://www.youtube- Disciplinary Conference on Web .com/yt/press/statistics.html. Acces- Accessibility - W4A ‘13, p.1. Avail- sed: 20th JANUARY 2016.