Millions of blind people cannot use travel sites effectively due to the significant access and usability barriers. Literature recognizes that travel sites lack accessibility and usability, but does not clarify the nature of interaction challenges blind users face. We conducted an exploratory field study with blind participants to examine where, how and why challenges arose in searching for flights on the Orbitz travel site. We employed verbal protocol analysis to generate an in–depth, contextually–situated and experiential understanding of participants’ interaction challenges. Results establish the feasibility and utility of our approach for an accurate evaluation of accessibility and usability of travel sites for blind users.
Travel sites such as Orbitz, Expedia and Priceline offer several benefits for the typical user. They serve as one–stop shop for meeting all travel–related needs (Law, et al., 2010). Users can search for services, prepare suitable itineraries and book reservations with only a few mouse clicks. Travel planning that once took hours now takes only a few hassle–free minutes (Mills, et al., 2006). Travel service providers offer exclusive discounted rates to the online customer (Lazar, et al., 2010). As a result, users make significant savings by booking online. Travel sites use a search engine technology that scours the Web for itineraries within a certain timeframe, service class, geographic location, and price range and find the best available deals, saving the traveler time and effort in research (Han and Mills, 2007). A side benefit of travel sites is bonus information such as weather updates, flight status, travel guides and tips (Carstens and Patterson, 2005). However, these benefits are not available if the user is blind due to the sight–centered design of the Web (Wentz, et al., 2011). This is the motivation behind this research.
The worldwide population with vision impairments is nearly 300 million (World Health Organization, 2013), who comprise a significant economic force. The net disposable income of blind Americans alone is worth US$175 billion (Lazar, et al., 2010). They engage in travel and tourism just as other members of society. They spend US$13 billion annually on travel–related services, including US$9.4 million on flights (Gutierrez, et al., 2005). They prefer travel sites over travel agents to take advantage of the attractive deals on offer and the additional options available. Literature informs that travel sites lack the accessibility and usability needed in non–visual interaction (Lazar, et al., 2011; Mills, et al., 2006; Han and Mills, 2007). However, it does not answer “What is the nature of interaction challenges that blind users face on a travel site?” An answer to this question is necessary to develop effective solutions that enhance the utility of these sites for blind travelers. Our research investigates this question with an aim to develop design principles on accessible and usable travel sites for the blind.
This paper presents a task–oriented and user–centered technique to evaluate the accessibility and usability of travel sites, and demonstrates its feasibility and utility for an in–depth, contextually–situated, experiential understanding of blind users’ interaction challenges. It reports the findings of an exploratory field study with blind participants. Think–aloud observations provided verbal evidence of their interaction experiences with the Orbitz site (www.orbitz.com). A combination of verbal protocol analysis (Ericsson and Simon, 1984) and an integrated problem–solving model (Babu, 2011) generated knowledge of where, how and why interaction challenges arose in performing a flight search. Results show that understanding flight search functionality, compiling a flight search query, executing this query, and evaluating search results were all problematic. It explains what specific aspects of these activities were challenging, what participants went through under these conditions, and what design elements were responsible. This knowledge was useful to make conjectures about potential design improvements to alleviate blind users’ flight search experiences on Orbitz.
Accessibility and usability are two critical factors that impact the utility of a Web site for users with disabilities (Babu and Singh, 2013). The evaluation technique presented here is versatile. It is useful to test the accessibility and usability of other travel site genres, in other task contexts, and for other disability types. Researchers in human–computer interaction, information science, the cognitive sciences and travel management could use this technique to accurately understand the interaction needs and challenges of travelers with disabilities. Developers and designers may use it to identify design errors in specific functionality of a travel site for assistive technology access. Travel organizations may use it to assess the utility of their Web sites for travelers with disabilities.
The rest of the paper is organized as follows. It begins with definition of key concepts and summary of relevant literature. It discusses the research design, including participants, material and procedure. What follows is the results and discussion. It ends with our concluding remarks and suggests possible directions for future research.
Here, we define key concepts and present a brief review of literature.
A blind user relies on screen reader (SR) software to interact with computers and the Web. SRs identify and interpret screen content through a synthetic voice (Di Blas, et al., 2004). Jaws, WindowEyes, VoiceOver and NVDA are commonly used SRs.
Web interaction for the blind is a listening activity. They perceive information on a Web page as continuous stream of audio. This perception becomes challenging when negotiating complex interface objects like menus and navigation bars (Babu and Singh, 2013). Typically, blind users jump elements with the Tab Key and explore content with Arrow keys (Buzzi, et al., 2010; Babu, 2011). Arriving on a new page, the SR announces the number of interface objects like links, images, and tables to contextualize the page structure (Babu, 2011). This non–visual interaction (NVI) is exacerbated in dynamic pages that use emergent technologies (Leuthold, et al., 2008). Users need extra mental effort for interpretation (Babu, 2011). Importantly, they are prone to developing inaccurate mental models without ‘clear view’ of the interface.
Accessibility and usability
Accessibility and usability are two related but distinct concepts. Accessibility allows users access to system functionality (Goodhue, 1988). For blind users, it is treated as a technical construct that allows screen readers the necessary access to interface elements (Leuthold, et al., 2008). Accessibility problems in travel sites prevent users’ access to necessary interface elements. Usability refers to how well a system conforms to users’ conceptualization of performing system–based tasks (Goodwin, 1987). It is a cognitive construct that depends on the task. Usability problems in travel sites prevent effective use of content and controls to meet objectives. An inaccessible site is unusable, but accessibility does not guarantee usability (Di Blas, et al., 2004). Both accessibility and usability are necessary for Web site utility (Babu and Singh, 2013).
The scant literature on Web accessibility and usability for the blind identifies the following challenges in non–visual interaction generalizable across Web sites:
Sequential. A Web page is rendered as a continuous stream of audio that lacks logical sections or segments. The user fails to recognize information conveyed through layout and organization (Salampasis, et al., 2005). When Web pages have a complex layout, the SR feedback becomes ambiguous (Lazar, et al., 2007).
Audio translation of on–screen text. What the user hears on a Web page is a mere translation of text content through synthetic speech, not a complete narration of information (Babu, 2011). There are no graphics in this rendition. Important cues embedded in color, images and videos that aid in navigation and interpretation are lost (Leuthold, et al., 2008).
Time–consuming. Inspecting a page at a glance is not possible. The user must listen to every bit of SR announcement carefully to locate goal–relevant information (Jones, et al., 2002). Browsing Web pages with high information density and innumerable hyperlinks is significantly cumbersome (Craven, 2003).
Unable to point and click. The predominant input method in non–visual interaction is keyboard–based. The user cannot interact with Web functionality that are operable through mouse alone (Chandrashekar, 2010).
Labelling is critical. The user perceives the affordance of an interface object such as hyperlink, button, checkbox and input field through its label alone. Improper labeling can cause significant confusion, frustration, and disorientation (Babu, 2011).
Things are not obvious. The user cannot tell when new content appears without full and continuous SR feedback. For example, she may be oblivious to a change in the page if the SR did not announce the download progress and the page composition (Babu and Singh, 2013). Input fields on a page become apparent only after receiving keyboard focus (Theofanos and Redish, 2003). Additionally, help features may go unnoticed. For example, users fail to recognize ‘Spelling Suggestions’ and ‘Related Searches’ offered in Google search (Sahib, et al., 2012).
Requires memorization. The SR has countless key commands for the wide range of functionality. To interact with a Web site effectively, the user must remember and use appropriate ones when needed. However, most users know or use only a handful of these (Theofanos and Redish, 2003).
Gets erratic in dynamic pages. The SR focus keeps shifting when interacting with pages where content is updated dynamically (Bigham, et al., 2007). This hampers the user’s ability to interact with relevant content effectively, efficiently, and in a satisfactory manner (Babu, 2011).
Cognitive overload. During interaction, the user must spend cognitive resources in trying to understand the browser, Web site, and SR simultaneously (Theofanos and Redish, 2003). Information duplicated across pages, such as Web site titles and navigational links, is announced repeatedly (Chandrashekar, 2010). SRs mispronounce many words (Theofanos and Redish, 2003), which creates comprehension problems. These factors contribute to cognitive overload in non–visual Web interaction (Millar, 1994; Thinus–Blanc and Gaunet, 1997).
Plug–in menace. Plug–ins such as PDF, Flash, Java applets, and Active X or the files they try to access are often incompatible with SRs (Lazar, et al., 2007). These plug–ins may also conflict with the speech engine, rendering the SR speechless (Babu, 2013). This situation is comparable to the ‘blue screen of death’.
These challenges may hinder a blind user’s ability to utilize travel sites effectively.
Accessibility and usability of travel Web sites
Very scant research investigates the accessibility of travel Web sites for the blind (Mills, et al., 2006; Han and Mills, 2007; Lazar, et al., 2010). It informs us that interacting with a travel Web site using an SR can be significantly challenging. For example, graphics such as pictures, virtual tours, and videos common on these sites present two potential problems for an SR user (Han and Mills, 2007). First, all information embedded in these graphics is lost if they are not marked appropriately. Second, they can temporarily disable audio feedback by causing an SR malfunction. Travel sites display an abundance of information such as schedules, options, and prices, interspersed with intrusive advertisements (Lazar, et al., 2010). This significantly hampers the browsing experiences of blind users, making information retrieval difficult or impossible. Travel Web sites often use Java and Macromedia files to embed links and menu options that do not support SR interaction (Han and Mills, 2007). This prevents access to such links and menu options for blind users. Flash can create two additional problems. It can trigger the SR to repeatedly read the top part of a Web page without reading the bottom part (BBC, 2013). Second, content updates cause the SR to return to the top of the page (Smith, et al., 2004). Two disruptive features of travel Web sites are the ads and pop–ups that do not interoperate with SRs effectively. This prevents blind users from reading Web pages properly. Actually, the SR treats a pop–up as a new Web page. This creates confusion and disorientation for blind users (Han and Mills, 2007). Hyperlinks without Alt–Text are another source of problems on travel Web sites; the SR announces their availability as “link, link, link” (Han and Mills, 2006). Another source of problems for blind users on travel Web sites is the online form. Such forms, when designed without any information assistance can create confusion for blind users (Babu, 2013). Locating the correct line and placing the cursor focus to fill out a form field can be challenging while making reservations on a travel Web site (Han and Mills, 2007). However, it is unclear how these challenges hinder blind users’ ability to perform basic travel–related functions such as booking a flight. In summary, literature recognizes that travel sites present accessibility problems in SR interaction, but does not clarify where, how and why blind users face challenges. This missing knowledge is needed to enhance the utility of travel sites for blind users.
Prior research (Babu, et al., 2010) developed a user–centered and task–oriented technique to examine blind users Web interactions. Task–orientation examines tasks and goals of interaction for a contextually–situated understanding of accessibility and usability problems in deriving intended systems utility. User–centeredness examines the needs, abilities, and challenges of blind users in systems interaction. It provides an understanding of users’ experiences in navigating systems accessibility and usability problems in performing tasks. We adapt this technique to study blind users’ interaction challenges on travel sites while developing in–depth, contextually–situated, experiential knowledge to inform improved design strategies for travel sites.
Participants included five individuals with diagnosed blindness (mean age = 21 years) residing in the Greater Milwaukee area at the time of the study. While three reported no light perception, the other two reported some light perception. Each reported using an SR for at least five years.
Material included a flight search task environment, an observation study protocol, and a laptop equipped with an SR and the Internet. The flight search task environment comprised multiple pages of the Orbitz online travel agency (www.orbitz.com) supporting five subtasks:
- Visit the travel Web site;
- Understand the flight search functionality;
- Create a flight search query;
- Execute the flight search query;
- Evaluate the flight search results.
The observation study protocol included a MS Word document describing the research objective and think–aloud method. It included a practice think–aloud task and instructions for completing the flight search task on Orbitz. This document was available on the home screen of the study laptop.
We distributed a flier to the National Federation of the Blind Milwaukee Chapter seeking volunteers as research participants. Interested volunteers e–mailed back to schedule study sessions at a convenient place and time. Each session commenced with the participant providing informed consent to participate. She then logged on to the study laptop and familiarized herself with the keyboard layout and the screen–reader version. She reviewed the study protocol document, and practiced thinking aloud for a simple search task on Google. She then visited the Orbitz Web site and performed the flight search task while thinking aloud. We audio–recorded her verbalization.
We performed verbal protocol analysis (Ericsson and Simon, 1984) on this verbal data. This commenced with transcription and segmentation. Each segment was analyzed for occurrence of Dissonance or Failure (Babu, 2011). Dissonance is a situation where a participant faced comprehension problems due to inadequate system feedback. Failure is a situation where actions yielded unexpected outcomes due to which participant could not operate something. Dissonance and failure represent interaction challenges for the blind user. Such interaction challenges are associated with design errors in the travel site. We synthesized the knowledge about dissonance and failure to develop an in–depth, contextually–situated, experiential understanding of accessibility and usability problems in conducting a flight search on Orbitz.
Analysis shows that four out of the five sub–tasks comprising a flight search on Orbitz were problematic for participants. It identified multiple interaction challenges that participants faced in each problematic subtask. For each interaction challenge, it provided an in–depth, contextually–situated, experiential understanding of their difficulties. In the following, we present results and supportive evidence pertinent to one interaction challenge per sub–task due to space constraints. Evidence comprises verbal protocols relevant to an episode of an interaction challenge, including participant utterances, SR announcements (enclosed within angular brackets), and system–generated non–verbal sounds (enclosed within square brackets).
Problem understanding flight search functionality
Analysis shows that participants faced problems understanding the flight search functionality of Orbitz. A major challenge was the dissonance regarding the need for dates in searching for flights. The dissonance stems from confusion about the utility of date fields in the flight search form.
Figure 1: Screenshot of the flight search form, including date fields.
The following segment of a participant’s transcript serves as evidence of this problem.
<Tab. mm slash dd slash yyyy.>
Oh! It’s probably asking for a birth date, month–date–year. Let me go down. Tab.
<Combo box, edit. Select, edit. Auto complete. Select, mm slash dd slash yyyy. Combo box edit, time. Collapse>
From what I am seeing here, it’s asking for your birth date, month, year. Pretty much what they would ask at an airport.
<From city name or airport, edit. Auto complete. Selected. Combo box edit. Edit, selected. Auto complete. Selected. mm slash dd slash yyyy>
I am confused about what to do with these date fields. It simply says some date has to go here. I am not sure whether it’s asking for my birth date or date of journey.
The above evidence demonstrates the participant’s difficulty understanding the purpose of the date fields in the flight search form. Although she could tell from the instruction on the date format that the input field was asking for a date, she did not have any clue about what date she must indicate. A lack of descriptive and meaningful caption for the input field communicating its utility for flight search contributes to this problem. It leaves the purpose of the date field open for interpretation. Consequently, the participant misinterpreted it to be the date of birth. Such misinterpretation contributes to an erroneous mental model of the flight search functionality. Such an erroneous mental model is likely to misguide blind users in searching for flights on Orbitz.
Problem creating flight search query
Analysis shows that participants faced problems compiling a flight search query. A major challenge was dissonance about the location in the flight search form. The dissonance results from being unable to determine what values are needed in the input fields at keyboard focus.
Figure 2: Screenshot of flight search form with keyboard focus on the “From” field.
The following segment from a participant’s transcript serves as evidence of this problem.
Right now, I am in an edit field. Probably it is asking what my current location is. I am not sure. Let me verify. I’m going to push Escape.
<Escape. Virtual pc>
And make sure I’m in the correct edit box. I am going up. Up Arrow.
<Link airport. From city name or airport>
Now Down Arrow.
I want to double check. Up Arrow.
<From city name or airport.>
Again Down Arrow.
Okay it says “from city name or airport” and then “edit”. I’m going to type in just Austin.
The above evidence demonstrates the participant’s difficulty determining what information to enter in the input field at keyboard focus. Although she was able to tell from the SR announcement “Edit” and the accompanying beep that she was currently placed at an input field, she did not find additional cues regarding what input to provide. This leaves the current location information relative to the search form an unknown. This contributed to an erroneous mental model of creating a flight search query that creates disorientation in navigating through the form’s input fields. To overcome this disorientation and dissonance, the participant surveys the surrounding area multiple times in search of this missing information. Only after spending four to five minutes, making three rounds of survey, and processing additional information could she understand the input field was asking for the originating airport for her flight. A lack of implicit labels for the airport input fields and contextual cues in the flight search form were responsible for this problem. The result of this design problem was extra time and effort in search of contextual cues and additional cognitive resources interpreting the input needed.
Problem executing flight search query
Analysis shows that participants faced problems executing a flight search query. A major challenge was failing to verify that the search query executed effectively. The failure results from the inability to perceive the activation of the “Search Flights” button.
Figure 3: Screenshot of the “Search Flights” button below the flight search from.
The following segments from a participant’s transcript act as evidence of this problem.
<Search Flights button>
“Search Flights button”. I’m going to click here to let it find the flights.
<Enter. New browser window. Blank>
It says “new browser window.”
<Trip dot com away marketing slash trip.>
I don’t know what that means.
Hurry up! I don’t know what’s going on here.
The above evidence demonstrates the participant’s inability to perceive if the “Search Flights” button activated successfully. Although she could tell from SR announcements such as “New browser window” and “7%” that a different Web page was loading, she did not receive additional feedback to verify that the search results page was available. Typically, blind users recognize a new Web page from SR announcement of the page title or the page composition such as “Page has 3 frames, 5 headings, and 52 links” (Babu and Singh, 2013). The inability to perceive the search results page meant the information about successful activation of the “Search Flights” button remained undiscovered. This gives rise to an erroneous mental model for executing a flight search query. To overcome this uncertainty and dissonance, the participant surveys the surrounding area in search of this missing information. However, she fails to locate any relevant information. In fact, irrelevant information such as “Trip dot com away marketing” further aggravates the uncertainty surrounding the execution of the flight search query. A lack of full and continuous feedback for activation of the “Search Flights” button was responsible for this problem. The result of this design problem was extra time and effort spent in search of clues about the outcome of the flight search query and additional cognitive resources interpreting the irrelevant information.
Problem evaluating flight search results
Analysis shows that participants faced problems evaluating the flight search results on Orbitz. A major challenge was dissonance regarding the selection of nearby airports needed for evaluation of flight search results. The dissonance stems from the inability to understand the relevance of “include airports within 80 miles” in the flight search results page.
Figure 4: Screenshot of the check box labeled “Include airport within 80 miles.”
The following segment of a participant’s transcript serves as evidence of this problem.
<Include airport within 80 miles checkbox unchecked. Leave. Edit text. Auto complete. Selected 06. Combo box>
It’s asking me “Do you like airport within 80 miles?” And it shows a button and a combo box. I’m not sure why it is asking that question again. I wonder if this is the search results page? It certainly didn’t say anything. I’m going to check if the search results are already out.
The above evidence demonstrates the participant’s difficulty understanding why he was asked to include airports within 80 miles in the flight search results page. Although the meaning of the question itself was unambiguous, the participant did not receive any feedback about its utility for evaluating the flight search results. What’s worse, he received no direct feedback as to what the outcome of the flight search query was. In practice, there could possibly be three outcomes: success — providing a list of flight options; failure — generating a message that no flights are available for the specified date for the specified airports; and error — generating a message about errors in creating the flight search query. This leaves the outcome of the flight search query undiscovered. This gives rise to an erroneous mental model of the outcome of the flight search query that misguides the participant. A lack of contextual cues pertinent to the outcome of the flight search query was responsible for the problem. The result of this design problem was extra time and effort in search of clues about the outcome of the flight search query and additional cognitive resources interpreting the question “include airports within 80 miles”.
Verbal protocol analysis generated an in–depth, contextually–situated, experiential knowledge of interaction problems that participants faced on Orbitz. It establishes that performing a flight search using a typical travel Web site by listening to a screen reader is difficult and cumbersome. Problems arise in effectively performing four of the five flight search activities — understanding the Web site’s flight search functionality, creating a flight search query, executing the flight search query, and evaluating the flight search results.
A major problem of blind users in understanding the flight search functionality is the dissonance about the dates needed for a flight search. From the feedback available on Orbitz, they may not understand what dates to specify in the date input fields. For instance, one of our participants assumed this to be her date of birth. The basis of this assumption apparently was a prior experience at an airport where she probably had to prove that she was old enough to travel independently. This assumption in a void of clear direction contributed to an erroneous mental model of the flight search functionality according to which one must disclose her date of birth in performing a flight search. Such erroneous mental models mislead users in attributing an incorrect purpose to a date input field. Based on principles of user–centered design (Nornan, 2002), this user problem is attributable to a design error in flight search form. Specifically, the lack of descriptive and meaningful captions for the date input fields prevent screen reader users effectively understand their purposes or information requested. The implication of this design error is extra time and effort understanding the flight search functionality, and errors in performing a flight search. The dissonance about date specification in flight search could potentially reduce with more descriptive and meaningful captions for date fields that unambiguously instruct users to indicate the dates of onward and return trips. Possibly, this will make the flight search functionality on Orbitz more intuitive, and help blind users avoid errors in date inputs.
A major problem of blind users in compiling flight search query is the dissonance about their location in the flight search form. From the feedback available on Orbitz, they may not recognize an input field at keyboard focus or tell what value is solicited. For example, our participants could not recognize the airport input field when it received keyboard focus. They explored the surrounding area looking for identifying information. This created difficulty ‘seeing’ the current location relative to the overall search form. This contributed to an erroneous mental model of compiling a flight search query. Such erroneous mental models could create disorientation in the minds of users while navigating through the flight search form. According to principles of user–centered design (Norman, 2002), this user problem is attributable to a design error in flight search form. Specifically, the absence of any implicit label for the airport input fields or contextual cues in areas surrounding these prevent screen reader users readily tell what is at current keyboard focus or what information is solicited. The implication of this design error is extra time and effort in search of contextual cues, and additional cognitive resources determining what input is needed. The dissonance about current location relative to the flight search form could potentially reduce with implicit labels that unambiguously ask the user for the departing and arriving airports respectively. This could possibly make current location information relative to the overall search form apparent, and help blind users avoid errors in compiling flight search queries.
A major problem for blind users in executing a flight search query is the failure to verify the submission of the completed flight search form. From the feedback available on Orbitz, they may be unable to tell if the “Search Flights” button activated. For example, our participants did not receive adequate feedback about the availability of the search results page after they hit Enter on the Search Flights button. In the absence of any announcement of the page title or the page composition, they could not tell if their flight search form was submitted. This contributed to an erroneous mental model for submitting a flight search form. Such erroneous mental models could give rise to uncertainty and ambiguity in executing flight search query. According to principles of user–centered design (Norman, 2002), this user problem is attributable to a design error in the flight search functionality of Orbitz. Specifically, the incomplete and inconsistent feedback following the activation of the “Search Flights” button prevent screen reader users from detecting the availability of flight options or error message in the subsequent page. The implication of this design error is extra time and effort searching for clues about the outcome of the flight search, and additional cognitive resources evaluating the relevance of every bit of information. The failure to verify the execution of the flight search query could potentially reduce with full and continuous feedback for activation of the “Search Flights” button, such as announcing the search results page title. This could possibly make the submission of the completed flight search form obvious, and help blind users avoid errors in executing flight search queries.
A major problem for blind users in evaluating flight search results is the dissonance on selection of nearby airports. From the information available on Orbitz search results page, they may not understand why they are required to pick another airport within a certain distance. For example, our participants failed to understand the relevance of “include airports within 80 miles” in the context of flight search results. The lack of further explanation as to why nearby airports be included, and any clue regarding the outcome of the flight search kept participants in the dark in regards to the search results. This contributed to an erroneous mental model for the outcome of flight search query. Such erroneous mental models could give rise to confusion and ambiguity in evaluating flight search results. According to principles of user–centered design (Norman, 2002), this user problem is attributable to a design error in the flight search functionality of Orbitz. Specifically, the obscurity of contextual cues relevant to the outcome of the flight search query made it difficult or impossible for screen reader users to locate the search results. The implication of this design error is extra time and effort searching for clues about the outcome of the flight search query and additional cognitive resources understanding the purpose of the question ‘include airports within 80 miles?’ The dissonance regarding the selection of nearby airports could potentially reduce with unambiguous feedback about the outcome of a flight search query — error, failure or success. This could help blind users readily evaluate the flight search results and rectify any error committed.
Prior research informs that travel sites lack the accessibility and usability needed by screen reader users. Our results illustrate the nature of problems blind users face in performing basic travel–related functions over a site. We demonstrate how to generate a better understanding of the needs and challenges of blind users in interacting with travel sites. Additionally, we demonstrate how to map a user problem to an error in the design of a travel site functionality. Such design errors help identify areas of improvement in a travel site accessible and usable to blind customers. Based on an understanding of how these errors impact blind users’ ability to search for a flight, we suggest feasible design modifications to potentially improve accessibility and usability of the flight search functionality. These suggested design improvements must undergo experimental validation before they could be treated as design principles. Table 1 summarizes our results and design suggestions:
Table 1: Snapshot of results and design suggestions. User problem Responsible design error Potential design solution Dissonance about dates needed in flight search Date field lacks descriptive and meaningful caption Date field include caption that solicit onward or return date Dissonance about location in flight search form Airport field lacks implicit label Airport field includes implicit label that solicits name of departing or arriving airport Failure to verify submission of flight search form Activation of “Search Flights” button generates incomplete and inconsistent feedback Full and continuous feedback of search results page following activation of “Search Flights” button Dissonance regarding selection of nearby airports Contextual cues relevant to outcome of flight search query are obscured Unambiguous feedback about outcome of flight search query — error, failure or success
The design errors identified here reflect some degree of ignorance among developers and designers about needs of blind users. Such ignorance is wide spread due to a paucity of literature on blind users’ Web interaction experiences. As the flight search task environment investigated here is more or less typical of what we observe across the travel industry, we can expect design errors such as improper labelling of input fields and inadequate cues on availability of search results in other travel sites that support flight bookings. Our on–going research is conducting a wide–scale investigation of blind users’ interaction experiences with multiple other travel sites. It will identify a comprehensive list of design errors and corresponding design recommendations. It will validate these design recommendations through an experiment. The output will be a checklist for travel agencies to offer an accessible and usable travel site for blind customers.
People who are blind spend billions out of their disposable income on travel and tourism (Gutierrez, et al., 2005). They are as likely to visit travel Web sites as their sighted counterparts. In fact, they prefer a travel Web site over a travel agency to benefit from the additional options available and lower prices (Han and Mills, 2007). However, prior research (Lazar, et al., 2011; Han and Mills, 2007; Williams and Rattray, 2005a; 2005b) show that 90 percent of travel Web sites have accessibility problems that hamper screen–reader interaction. These problems prevent blind users from using these Web sites effectively (Lewis and Klauber, 2002; Maeda, et al., 2004). When interaction challenges arise in basic tasks such as flight search, they develop negative perceptions about the travel site and switch loyalty to a more accessible alternative (Law and Hsu, 2005). Hence, it is important that travel organizations improve the accessibility and usability of their Web sites so that blind users can search for travel–related products, make travel reservations, and leave feedback effectively, efficiently and in a satisfactory manner. They need a fresh evaluation of these Web sites with a proper understanding of blind users’ needs and challenges (Law and Hsu, 2005). This paper presents a feasible technique to develop an in–depth, contextually–situated, experiential understanding of the needs and challenges of blind users in interacting with travel Web sites, and use this as the basis to trace accessibility and usability problems in interface design.
Results of our exploratory field study establish that travel Web site continue to present access and usability barriers in screen reader interaction. The technique combines task analysis and verbal protocol analysis to explain the nature of a problem at an elementary level, and map it to the responsible design elements. This guided the identification of feasible design improvements that can potentially reduce or eliminate blind users’ problems. Subject to validation, such design improvements form the basis of design principles on accessibility and usability of travel Web sites for blind users. Results provide an in–depth, contextually–situated, experiential understanding of interaction challenges in performing a flight search. The technique is useful to examine other task contexts (e.g., booking flights, searching for hotel accommodations, etc.) and other travel Web sites (e.g., Web sites of airlines and hotels, expedia.com, priceline.com, kayak.com, etc). Our on–going research is performing a wide–scale evaluation involving different travel Web sites, different task contexts, using a more diverse sample of blind users. We expect this will generate a list of task-specific design recommendations to improve accessibility and usability. We will vet these design recommendations from designers of travel Web sites and existing standards on Web accessibility and usability. We will validate these recommendations in an experiment using a prototype travel site. The outcome will be a set of design principles on accessible and usable travel sites for blind users.
About the author
Rakesh Babu is an assistant professor in the School of Information Studies at the University of Wisconsin–Milwaukee. He is blind and has strong motivation to undertake and promote research topics relevant to the empowerment of the blind in the information society. His research expertise includes Web accessibility and usability, human–centered computing, cognitive models, online education, social computing, travel & tourism management, and verbal protocol analysis. He has multiple prior and on–going externally–funded research projects sponsored by the National Science Foundation, European Research Council, and Research Council of Norway. His research appears in multiple peer–reviewed journals and conference proceedings. He is a regular reviewer for multiple journals and international conferences.
E–mail: babu [at] uwm [dot] edu
Rakesh Babu, 2013. “Developing more accurate competence models for improved e–learning for the blind,” Journal of Information Science and Technology, volume 9, number 2.
Rakesh Babu, 2011. “Developing an understanding of the accessibility and usability problems blind students face in Web–enhanced instruction environments,” Ph.D. dissertation at the University of North Carolina at Greensboro; see http://libres.uncg.edu/ir/uncg/f/Babu_uncg_0154D_10785.pdf, accessed 10 December 2013.
Rakesh Babu and Rahul Singh, 2013. “Enhancing learning management systems utility for blind students: A task–oriented, user–centered, multimethod evaluation technique.” Journal of Information Technology and Education: Research, volume 12, at http://www.jite.org/documents/Vol12/JITEv12ResearchBabu001-032.pdf, accessed 10 July 2013.
Rakesh Babu, Rahul Singh, and Jai Ganesh, 2010. “Understanding blind users’ Web accessibility and usability problems,” AIS Transactions on Human–Computer Interaction, volume 2, number 3, http://aisel.aisnet.org/thci/vol2/iss3/1/, accessed 10 December 2013.
BBC, 2013. “Multimedia accessibility standards v1.0,” at http://www.bbc.co.uk/guidelines/futuremedia/accessibility/multimedia.shtml, accessed 2 July 2013.
Jeffrey P. Bigham, Anna C. Cavender, Jeremy T. Brudvik, Jacob O. Wobbrock, and Richard E. Lander, 2007. “WebinSitu: A comparative analysis of blind and sighted browsing behavior,” Assets ’07: Proceedings of the 9th International ACM SIGACCESS Conference on Computers and Accessibility, pp. 50–58.
http://dx.doi.org/10.1145/1296843.1296854, accessed 10 December 2013.
Maria Claudia Buzzi, Marina Buzzi, Barbara Leporini, and Fahim Akhter, 2010. “Is Facebook really ‘open’ to all?” 2010 IEEE International Symposium on Technology and Society (ISTAS), pp. 327–336.
http://dx.doi.org/10.1109/ISTAS.2010.5514621, accessed 10 December 2013.
Deborah Sater Carstens and Pauline Patterson, 2005. “Usability study of travel Web sites,” Journal of Usability Studies, volume 1, number 1, pp. 47–61, and at http://www.usabilityprofessionals.org/upa_publications/jus/2005_november/travel.html, accessed 10 December 2013.
Sambhavi Chandrashekar, 2010. “Is hearing believing? Perception of online information credibility by screen reader users who are blind or visually impaired,” doctoral dissertation, University of Toronto, at https://tspace.library.utoronto.ca/handle/1807/26157, accessed 10 December 2013.
Jenny Craven, 2003. “Accessibility and usability of Web sites,” Library and Information Update, volume 2, number 5, pp. 46–47.
Nicoletta Di Blas, Paolo Paolini, and Marco Speroni, 2004. “‘Usable accessibility’ to the Web for blind users,” Proceedings of the 8th ERCIM Workshop on User Interfaces for All, pp. 28–29, and at http://ui4all.ics.forth.gr/workshop2004/files/ui4all_proceedings/adjunct/accessibility/109.pdf, accessed 10 December 2013.
K. Anders Ericsson and Herbert A. Simon, 1984. Protocol analysis: Verbal reports as data. Cambridge, Mass.: MIT Press.
Dale L. Goodhue, 1988. “I/S attitudes: Toward theoretical and definitional clarity,” ACM SIGMIS Database, volume 19, numbers 3–4, pp. 6–15.
http://dx.doi.org/10.1145/65766.65768, accessed 10 December 2013.
Nancy C. Goodwin, 1987. “Functionality and usability,” Communications of the ACM, volume 30, number 3, pp. 229–233.
http://dx.doi.org/10.1145/214748.214758, accessed 10 December 2013.
Charletta F. Gutierrez, Constantine Loucopoulos, and Roger W. Reinsch, 2005. “Disability–accessibility of airlines’ Web sites for US reservations online,” Journal of Air Transport Management, volume 11, number 4, pp. 239–247.
http://dx.doi.org/10.1016/j.jairtraman.2005.01.006, accessed 10 December 2013.
Jee–Hee Han and Juline Mills, 2007. “Are travel Web sites meeting the needs of the visually impaired?” Information Technology & Tourism, volume 9, number 2, pp. 99–113.
http://dx.doi.org/10.3727/109830507781367401, accessed 10 December 2013.
Jee–Hee Han and Juline Mills, 2006. “The mutual designing of travel Web sites: Perceptions of the visually impaired,” Information and Communication Technologies in Tourism 2006: Proceedings of the International Conference in Lausanne, Switzerland, pp. 147–157.
http://dx.doi.org/10.1007/3-211-32710-X_22, accessed 10 December 2013.
Keith S. Jones, J. Shawn Farris, Peter D. Elgin, Brent A. Anders, and Brian R. Johnson, 2002. “Examining information searching on the World Wide Web with a screen–reader: A verbal protocol analysis,” Proceedings of the Human Factors and Ergonomics Society Annual Meeting, volume 46, number 14, pp. 1,310–1,314.
http://dx.doi.org/10.1177/154193120204601415, accessed 10 December 2013.
Rob Law and Cathy H.C. Hsu, 2005. “Customers’ perceptions on the importance of hotel Web site dimensions and attributes,” International Journal of Contemporary Hospitality Management, volume 17, number 6, pp. 493–503.
http://dx.doi.org/10.1108/09596110510612130, accessed 10 December 2013.
Rob Law, Shanshan Qi, and Dimitrios Buhalis, 2010. “Progress in tourism management: A review of Web site evaluation in tourism research,” Tourism Management, volume 31, number 3, pp. 297–313.
http://dx.doi.org/10.1108/09596110510612130, accessed 10 December 2013.
Jonathan Lazar, Brian Wentz, Dana Biggers, Jason Delair, Matthew Donnelly, Eludoyin Kashim, Andrew Henin, James Markakis, Angela Matos, Andrew McNicol, John Nixon III, Richard Osborne, Tatyana Postnova, Jeevakumar Raja, Roland Roberts, Harry Serra III, Vasilios Sfakianoudis, Vincent Tyler, and Junhyeok Yun, 2011. “Societal inclusion: Evaluating the accessibility of job placement and travel Web sites,” Proceedings of the INCLUDE 2011 Conference, at http://www.hhc.rca.ac.uk/3845/all/1/proceedings.aspx, accessed 10 December 2013.
Jonathan Lazar, Paul T. Jaeger, Anthony Adams, Anthony Angelozzi, John Manohar, James Marciniak, Justin Murphy, Pouria Norasteh, Charles Olsen, Evangelos Poneres, Tiffany Scott, Naresh Vaidya, and James Walsh, 2010. “Up in the air: Are airlines following the new DOT rules on equal pricing for people with disabilities when Web sites are inaccessible?” Government Information Quarterly, volume 27, number 4, pp. 329–336.
http://dx.doi.org/10.1016/j.giq.2010.04.005, accessed 10 December 2013.
Jonathan Lazar, Aaron Allen, Jason Kleinman, and Chris Malarkey, 2007. “What frustrates screen reader users on the Web: A study of 100 blind users,” International Journal of Human–Computer Interaction, volume 22, number 3, pp. 247–269.
Stefan Leuthold, Javier A. Bargas–Avila, and Klaus Opwis, 2008. “Beyond Web content accessibility guidelines: Design of enhanced text user interfaces for blind Internet users,” International Journal of Human–Computer Studies, volume 66, number 4, pp. 257–270.
http://dx.doi.org/10.1016/j.ijhcs.2007.10.006, accessed 10 December 2013.
Valerie Lewis and Julie Klauber, 2002. “Inaccessible Web design from the perspective of a blind librarian,” Library Hi Tech, volume 20, number 2, pp. 137–140.
http://dx.doi.org/10.1108/07378830210432499, accessed 10 December 2013.
Junji Maeda, Kentarou Fukuda, Hironobu Takagi, and Chieko Asakawa, 2004. “Web accessibility technology at the IBM Tokyo Research Laboratory,” IBM Journal of Research and Development, volume 48, numbers 5–6, pp. 735–749.
Susanna Millar, 1994. Understanding and representing space: Theory and evidence from studies with blind and sighted children. Oxford: Oxford University Press.
Juline E. Mills, Jee–Hee Han, and Joan Marie Clay, 2008. “Accessibility of hospitality and tourism Web sites: A challenge for visually impaired persons,” Cornell Hospitality Quarterly, volume 49, number 1, pp. 28–41.
http://dx.doi.org/10.1177/1938965507311499, accessed 10 December 2013.
Nuzhah Gooda Sahib, Anastasios Tombros, and Tony Stockman, 2012. “A comparative analysis of the information–seeking behavior of visually impaired and sighted searchers,” Journal of the American Society for Information Science and Technology, volume 63, number 2, pp. 377–391.
http://dx.doi.org/10.1002/asi.21696, accessed 10 December 2013.
Michail Salampasis, Christos Kouroupetroglou, and Athenasios Manitsaris, 2005. “Semantically enhanced browsing for blind people in the WWW,” HYPERTEXT ’05: Proceedings of the Sixteenth ACM Conference on Hypertext and Hypermedia, pp. 32–34.
http://dx.doi.org/10.1145/1083356.1083363, accessed 10 December 2013.
Ann C. Smith, Justin S. Cook, Joan M. Francioni, Asif Hossain, Mohd Anwar, and M. Fayezur Rahman, 2004. “Nonvisual tool for navigating hierarchical structures,” Assets ’04: Proceedings of the 6th International ACM SIGACCESS Conference on Computers and Accessibility, pp. 133–139, and ACM SIGACCESS Accessibility and Computing, numbers 77-78, pp. 133–139.
http://dx.doi.org/10.1145/1029014.1028654, accessed 10 December 2013.
Mary Frances Theofanos and Janice Redish, 2003. “Guidelines for accessible and usable Web sites: Observing users who work with screen readers,” Interactions, volume 10, number 6, pp. 38–51.
Catherine Thinus–Blanc and Florence Gaunet, 1997. “Representation of space in blind persons: Vision as a spatial sense?” Psychological Bulletin, volume 121, number 1, pp. 20–42.
http://dx.doi.org/10.1037/0033-2909.121.1.20, accessed 10 December 2013.
Brian Wentz, Paul T. Jaeger, and Jonathan Lazar, 2011. “Retrofitting accessibility: The legal inequality of after–the–fact online access for persons with disabilities in the United States,” First Monday, volume 16, number 11, at http://firstmonday.org/article/view/3666/3077, accessed 10 July 2013.
Russell Williams and Rulzion Rattray, 2005a. “UK and US hotel Web content accessibility,” International Journal of Contemporary Hospitality Management, volume 17, number 1, pp. 78–87.
http://dx.doi.org/10.1108/09596110510577699, accessed 10 December 2013.
Russell Williams and Rulzion Rattray, 2005b. “UK hotel Web page accessibility for disabled and challenged users,” Tourism and Hospitality Research, volume 5, number 3, pp. 225–268.
http://dx.doi.org/10.1057/palgrave.thr.6040024, accessed 10 December 2013.
World Health Organization, 2013. “Visual impairment and blindness,” at http://www.who.int/mediacentre/factsheets/fs282/en/, accessed 3 March 2013.
Received 18 July 2013; revised 24 September 2013; accepted 25 September 2013.
Copyright © 2013, First Monday.
Copyright © 2013, Rakesh Babu. All Rights Reserved.
Understanding challenges in non–visual interaction with a travel site: An exploratory field study with blind users
by Rakesh Babu.
First Monday, Volume 18, Number 12 - 2 December 2013