Designing the Navigation to E-services Improving the Accessibility and Usability to Report a Crime or Loss at the Swedish Police Authority’s Website Master’s Thesis in Computer Science and Engineering MAJA ALBREKTSON & CAROLINE XIA Department of Computer Science and Engineering CHALMERS UNIVERSITY OF TECHNOLOGY UNIVERSITY OF GOTHENBURG Gothenburg, Sweden 2021 Master’s Thesis 2021 Designing the Navigation to E-services Improving the Accessibility and Usability to Report a Crime or Loss at the Swedish Police Authority’s Website MAJA ALBREKTSON & CAROLINE XIA Department of Computer Science and Engineering Chalmers University of Technology University of Gothenburg Gothenburg, Sweden 2021 Designing the Navigation to E-services Improving the Accessibility and Usability to Report a Crime or Loss at the Swedish Police Authority’s Website MAJA ALBREKTSON & CAROLINE XIA © MAJA ALBREKTSON & CAROLINE XIA, 2021. Supervisor: Sara Ljungblad, Department of Computer Science and Engineering Advisor: Nina Mujdzic, Swedish Police Authority Examiner: Staffan Björk, Department of Computer Science and Engineering Master’s Thesis 2021 Department of Computer Science and Engineering Interaction Design Division Chalmers University of Technology and University of Gothenburg SE-412 96 Gothenburg Telephone +46 31 772 1000 Cover: An illustration of the design concept when the user has answered all relevant questions and been presented with information about which service to use. Typeset in LATEX Gothenburg, Sweden 2021 iv Designing the Navigation to E-services Improving the Accessibility and Usability to Report a Crime or Loss at the Swedish Police Authority’s Website MAJA ALBREKTSON & CAROLINE XIA Department of Computer Science and Engineering Chalmers University of Technology and University of Gothenburg Abstract In the fast-paced digital world, authorities and organizations are moving towards developing and providing digital services on the internet. Regarding the Swedish Police Authority, citizens can use e-services to report crimes or losses. However, the authority has acknowledged that multiple citizens use the wrong e-service for their case and sometimes have trouble finding the correct one. The thesis aimed to explore how the web navigation to the e-services can be im- proved to help the citizens find the correct service for their circumstances. The investigation of the topic is based on a user-centered design process where 14 meth- ods have been executed. The methods included data gathering, development of requirements and exploration of a potential design concept. The results are presented in 15 design guidelines about improving the navigation to e-services regarding usability and accessibility. Furthermore, a design concept is presented that exemplifies how the guidelines can be explicitly applied to the Swedish Police Authority’s website. Keywords: user experience, user research, interaction design, design guidelines, design concept, data collection, usability, accessibility, computer science, e-services v Acknowledgements We would like to express our gratitude to all the people that made this thesis pos- sible. To our supervisors Sara Ljungblad and Nina Mujdzic for your never ending support and guidance. We can not emphasize enough how helpful it has been to know that you always had our backs. Thank you! To all the participants in our studies - this project would not have been possible without you. Thank you for your time, engagement and feedback. We would also like to dedicate a thanks to all the employees at the Swedish Police Authority, that welcomed us with open arms and made us feel comfortable. A spe- cial thanks to Daniel Svan who believed in us and gave us this opportunity. Thank you all for your support and interest in our project! Maja Albrektson & Caroline Xia Gothenburg, 2021 vii Contents 1 Introduction 3 1.1 Research Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2 Research Question . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3 Aim & Deliverables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.4 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2 Background 7 2.1 The Swedish Police Authority . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Stakeholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.3 Ethical Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4 Previous Work and the Thesis’s Contribution . . . . . . . . . . . . . 10 2.4.1 E-Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.4.2 Website Navigation . . . . . . . . . . . . . . . . . . . . . . . . 10 2.4.3 Accessibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.4.4 The Thesis’s Contribution . . . . . . . . . . . . . . . . . . . . 12 3 Theory 13 3.1 Website Navigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 3.1.1 Navigational Models . . . . . . . . . . . . . . . . . . . . . . . 13 3.1.2 Website Navigation - Design Guidelines . . . . . . . . . . . . . 16 3.1.3 Information Foraging . . . . . . . . . . . . . . . . . . . . . . . 16 3.2 Decision Making Theory . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.2.1 Heuristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.2.2 Biases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3 Design System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 3.4 Usability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.5 Inclusive Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.5.1 Disability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4 Methodology 27 4.1 Research Through Design . . . . . . . . . . . . . . . . . . . . . . . . 27 4.1.1 Wicked Problems . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.2 Remote Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.3 User Research Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.4 Design Thinking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4.4.1 Empathize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 ix Contents 4.4.2 Define . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.4.3 Ideate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 4.4.4 Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4.4.5 Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5 Planning 39 5.1 Project Initiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.2 Design Thinking Process . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.2.1 Empathize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.2.2 Define . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.2.3 Ideate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.2.4 Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 5.2.5 Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 5.3 Report Writing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 6 Execution and Process 45 6.1 Empathize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 6.1.1 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . 45 6.1.2 Simulation Exercise . . . . . . . . . . . . . . . . . . . . . . . . 45 6.1.3 Indirect Observation . . . . . . . . . . . . . . . . . . . . . . . 47 6.1.4 User Study (Eye Tracking and Questionnaire) . . . . . . . . . 49 6.1.5 Think Aloud with a Visually Impaired Person . . . . . . . . . 53 6.1.6 Expert Interviews . . . . . . . . . . . . . . . . . . . . . . . . . 55 6.2 Define . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 6.2.1 Affinity Diagramming . . . . . . . . . . . . . . . . . . . . . . 57 6.2.2 Behavioral Archetype . . . . . . . . . . . . . . . . . . . . . . . 62 6.2.3 Design Requirements . . . . . . . . . . . . . . . . . . . . . . . 64 6.3 Ideate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 6.3.1 Sketching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 6.3.2 Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 6.4 Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.4.1 Digital Prototyping (High Fidelity Prototyping) . . . . . . . . 71 6.5 Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.5.1 Usability Test with Experts from the Contact Center . . . . . 71 6.5.2 Usability Test with UX Designers at the Authority . . . . . . 73 6.5.3 Think Aloud & Interviews with Users . . . . . . . . . . . . . . 74 6.5.4 Usability Test with Accessibility Expert . . . . . . . . . . . . 77 7 Results 79 7.1 Design Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 7.1.1 Home Page & Middle Page . . . . . . . . . . . . . . . . . . . . 81 7.1.2 Understanding If an E-service is the Appropriate Choice . . . 83 7.1.3 Selecting a Crime or Loss . . . . . . . . . . . . . . . . . . . . 84 7.1.4 Clarifying the Crime or Loss . . . . . . . . . . . . . . . . . . . 85 7.1.5 The End of the Guide . . . . . . . . . . . . . . . . . . . . . . 86 7.2 Design Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 7.2.1 Design - Look & Feel . . . . . . . . . . . . . . . . . . . . . . . 89 x Contents 7.2.2 Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 7.2.3 Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 7.2.4 Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 7.2.5 Methods & Process . . . . . . . . . . . . . . . . . . . . . . . . 96 8 Discussion 97 8.1 Design Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 8.2 Design Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 8.3 Methodology Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 99 8.3.1 Simulation Exercise . . . . . . . . . . . . . . . . . . . . . . . . 99 8.3.2 Eye Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 8.3.3 Behavioral Archetypes . . . . . . . . . . . . . . . . . . . . . . 100 8.3.4 Usability Testing . . . . . . . . . . . . . . . . . . . . . . . . . 101 8.3.5 Different Methods of Evaluation . . . . . . . . . . . . . . . . . 101 8.3.6 Remote Testing with Users . . . . . . . . . . . . . . . . . . . . 102 8.4 Transferability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 8.5 Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 8.6 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 9 Conclusion 107 References 109 Appendix I A Conclusions, Simulation Exercise . . . . . . . . . . . . . . . . . . . . I B Consent Form, User Study . . . . . . . . . . . . . . . . . . . . . . . . II C Information Form, User Study . . . . . . . . . . . . . . . . . . . . . . III D Questionnaire Questions . . . . . . . . . . . . . . . . . . . . . . . . . IV E Interview Guide, Accessibility Expert . . . . . . . . . . . . . . . . . . IX F Interview Guide, UX Design Expert . . . . . . . . . . . . . . . . . . . X G Behavioral Archetype, the ’Good Enough’ Gale . . . . . . . . . . . . XI H Behavioral Archetype, the ’Careful’ Charlie . . . . . . . . . . . . . . . XII I Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIII J Usability Test with Experts from Contact Center . . . . . . . . . . . XX K Think Aloud & Interviews with Users . . . . . . . . . . . . . . . . . . XXI 1 Contents 2 1 Introduction In Sweden, all citizens who have been the victims of a crime, regardless of when it happens in life, need to contact the police. The Swedish Police Authority is a public service with the purpose to "reduce crime and increase public safety" (The Swedish Police, n.d.). To achieve this, effective communication with citizens is of the utmost importance. The digital development combined with new technology has made citizens accus- tomed to a digital world, thus creating higher expectations of e-services (electronic services) in the public sector (Bharosa et al., 2020). Furthermore, all public sectors in Sweden need to follow specific web accessibility guidelines (Infrastrukturdeparte- mentet RSED DF, 2018). To meet this need, public sectors around the world have realized that it should be easy to use their services without knowing all the bureau- cracies behind (OECD, 2017). Personalization and more ease of use are expected from the users, making digital technologies essential while simultaneously offering a reduced cost (Bharosa et al., 2020). Furthermore, it is becoming more common to create holistic solutions putting the need of the user and the feedback from the citizens in focus (OECD, 2017). Users want to easily find the information they are looking for. However, it is fre- quently reported that users have trouble with navigational tasks such as finding a specific site or content. In some contexts, these troubles might lead to the user using another site or service instead (Lazar et al., 2003). Regarding the Swedish Police Authority, such frustration might lead to more severe consequences, such as ignoring reporting a crime. That, in turn, might lead to other consequences, includ- ing skewed statistics, unresolved crimes, and dissatisfied citizens. Thus, the need to make efficient navigation with the user in focus is more critical than ever. 1.1 Research Problem To report some crimes and losses, the citizens can use e-services on the Swedish Police Authority’s website. However, the Police Authority has recognized that many reports are done using the wrong e-service (e.g. burglary instead of theft). In some cases, when the users are having trouble finding the correct one, they decide not to finish the report and call 114 14 for personal help instead. One example of how this 3 1. Introduction might happen and the potential consequences is presented in the scenario below. Kim was tired after a long day but decided to go down to the base- ment and pick out the Easter decorations. On arrival, Kim noticed that someone had tried to break in. The lock was destroyed, and all the decorations were gone. Kim felt very stressed and upset while return- ing to the apartment. After searching the internet and looking at the Swedish Police Authority website, Kim started to report the crime as ’Förlustanmälan’. However, the e-service asked for questions not related to the crime, and there was no option to report the broken lock as a consequence of the burglary. Is this the correct e-service to use? Kim felt stupid not being able to use the service but decided that it was more important to do it correctly and called the police instead. First, there was a telephone queue with a waiting time of 20 minutes. When Kim finally reached a person on the other side, the registrar needed to note everything manually while Kim patiently spelt the name and address letter by letter. When hanging up, Kim felt very frustrated that this matter occupied the entire evening. This fictional scenario is not unreasonable and might lead to multiple problems for the authority as well. The case will be delayed, the telephone line will be loaded, and resources that could be used for other things are occupied. In other scenarios, the user might continue to use the e-service even though it is incorrect. Hence, the user will answer questions that are irrelevant to the crime or loss they are reporting. Moreover, it is a growing need to create digital accessibility for everyone. DIGG - Myndigheten för digital förvaltning (n.d.) is an agency for digital government aimed to improve digital accessibility on all public websites, including the Swedish Police Authority. Therefore, it is crucial to make navigation towards the e-services accessible for all user groups and include a diversity of users that covers variation in capabilities, needs, and aspirations. The e-service team at Swedish Police Authority have suggested a potential solution to make people find the correct e-services. The solution is to incorporate a question guide before users access the different e-services. The question guide will collect rel- evant information from the users and, based on that, redirect the user to the correct e-service. The thesis will consider the findings and insights gathered throughout the process and ideate on all ideas to find the optimal solution. 1.2 Research Question This research problem has resulted in the following research question and sub- question: What should be considered when designing an accessible and usable web navigation to e-services? 4 1. Introduction The research question will be answered by a set of guidelines. Thus, the sub-question is: How could such guidelines be applied to the Swedish Police Authority’s website re- garding the e-services that are used to report a crime or loss? 1.3 Aim & Deliverables The project aims to improve the navigation to different e-services at the Swedish Police Authority’s website, focusing on the e-services that are used to report a crime or loss. The project will follow a user-centered design process called the design thinking process covering the phases: empathize, define, ideate, prototype and test. A crucial part of the process will be interacting with the users and experts in the field to investigate their perspective and create a seamless interaction that guides them toward the correct e-service. The purpose is to contribute with knowledge that gives the Swedish Police Authority a better understanding of improvements made to create user-friendly navigation, which involves understanding the users and their needs. To reach the project’s aim, the thesis will result in deliverables related to the re- search questions. The deliverables for the thesis are guidelines to consider when improving the navigation of a website with e-services and a high fidelity prototype that visualizes an example of how the guidelines can be implemented. 1.4 Limitations The thesis’s focus is to create improved navigation to the e-services that are used for reporting a crime or loss. Moreover, the improved navigation should be reusable, making it possible to develop it and include more e-services in the future. The design will be made for desktop devices and created in Swedish since that is the primary language for the website. The thesis also focuses on guiding users in the navigation towards the e-services used to report a crime or loss. To narrow it down, the thesis will not cover the navigation to other e-services on the site, e.g. the services that are used to apply for permission. Nor will there be any focus on how to best design the actual e-services. The findings from the study might apply to other authorities and organizations that need a similar solution. However, there might be a need for corrections to suit those specific circumstances. Due to the pandemic of Covid-19, no methods related to physical meetings will be conducted. Instead, the selection of methods will be adjusted and performed online. If not possible, alternative methods will be used. Furthermore, the thesis will not include the implementation of the prototype in 5 1. Introduction terms of code. The launching of a coded product would make the project heavily dependent on external parties, risking the deadline of the academic part. Thus, that part will be assigned to people within the IT department at the Swedish Police Authority. 6 2 Background This chapter aims to bring knowledge into the thesis’s background regarding the Swedish Police Authority and the stakeholders. Moreover, the chapter presents con- siderations about the ethical aspects. Finally, some related work within the domain is highlighted, and the thesis’s contribution to the academic world is presented. 2.1 The Swedish Police Authority The Swedish Police Authority is a one-council authority operated by the National Police Chief, responsible for the Police activities. The Swedish Police states that their mission is to reduce crime and increase public safety. The mission includes to "monitor public order and safety, conduct reconnaissance and carry out criminal investigations" (Polisen, n.d.-b). There are several ways to report a crime or a loss. Citizens can call in with the numbers: 112 for urgent help or 114 14 for other cases such as police reports, tip- offs or information. Other ways to report are visiting the police station, reporting via a police on-the-spot, or using the different e-services on the website. Some of the e-services are report of theft, credit card fraud, shoplifting, damage and parking damage (Polisen, 2020). Altogether, all the e-services receive about 370 000 case registrations per year. The e-services can be reached from the start page; see figure 2.1 for an overview of the structure today. The website of the Swedish Police Authority provides external information and con- sists of approximately 20 different e-services accessible to the citizens. An overview of today’s navigation to report a burglary is illustrated in figure 2.1. 7 2. Background 1. 2. 3. 4. Figure 2.1: Overview of today’s navigation to report a burglary. (1) The user selects the button ’I have been through something’, (2) In the list of e-services to report a crime or loss, the user selects the link to ’Report a burglary’, (3) The user selects the button ’Start e-service’ and (4) the e-service of reporting a burglary is presented. Authors’ own collage of screenshots from Polisen (n.d.-a). 2.2 Stakeholders The thesis has several stakeholders that are interested in or will be affected by its result. The two researchers Maja Albrektson and Caroline Xia, are enrolled 8 2. Background at Chalmers University of Technology and will perform the thesis as part of the academic institution program Interaction Design of Technologies. The university is interested in the thesis to receive an academic report that will contribute to research within the interaction design discipline. Furthermore, the supervisor from the university, Sara Ljungblad, will give academic support during the process of conducting the research and writing the report. The client for the project is the Swedish Police Authority in Gothenburg, primarily the IT department of e-services. The department is responsible for the e-services that interact with the general public. The department has provided a mentor for this project, Nina Mujdzic, who works as a Business Analyst and UX Designer in the assigned area. The expectation from the client is results that can contribute to knowledge on how to improve the navigation to their e-services. Other stakeholders for the thesis are the participants who will be recruited for data collection and evaluation. From their perspective, sensible information may be col- lected during the project. Because of that, the researchers will ensure informed consent and comply with ethical guidelines and laws that regulate the ethical re- quirements for the research process. The target group could also be seen as stakeholders. For this project, the target group consist of mainly every citizen that in some time, either in the future or now, needs to report a crime using an e-service at the Swedish Police Authority website. Therefore, they will be affected by the result of this project. Hereafter, the client the Swedish Police Authority, will be referred to as the authority in this thesis. 2.3 Ethical Considerations Ethical considerations play an essential role when conducting research. The re- searchers have a responsibility to follow the ethical guidelines towards all who are participating in the research and towards all stakeholders who may be affected by the results of it. The end-users of the project are every citizen that some time, either in the future or now, need to report a crime to the authority. It is a wide range of people, and there will be a need for data gathering from the users to practice user-centered design and adjust the system after their needs. The data gathering will be made with informed consent, anonymity and confidentiality by using a consent form or similar to comply with the ethical guidelines and laws that regulate and set ethical requirements for the research process (Vetenskapsrådet, 2017). That involves informing the participants about what data will be collected, how it will be stored and how it will be used. It is also crucial to know data collection regulations, for example, GDPR (The General Data Protection Regulation). The researchers will consider the potential ethical issues during the whole design pro- 9 2. Background cess and do their best to minimize their impact. Potential recordings such as sound and video will be destroyed after the study to ensure the participants’ anonymity. All participation will be voluntary. 2.4 Previous Work and the Thesis’s Contribution This section defines the concept of e-services, and presents opportunities as well as challenges regarding these. Furthermore, previous work within the area of website navigation is presented. The section also brings up information about the accessi- bility perceptive and concludes with the thesis’s contribution. 2.4.1 E-Services E-services can be defined as web-based services or interactive services that are pro- vided by the use of information and communication technologies (ICTs) (Rowley, 2006). Furthermore, Hoffman and Bateson (1997) states that e-services can be de- fined as: "E-service is deeds, efforts or performances whose delivery is mediated by information technology (including the Web, information kiosks and mobile devices). Such e-service includes the service element of e-tailing, customer support and service, and service delivery.". It is a broad domain including both e-businesses which acts in the private sector, and e-government which means that government provide e-services to citizens (Rowley, 2006). There is no face-to-face interaction between the customer and the organi- zation. Instead, the interaction happens through the technology, such as a website (Rowley, 2006). The emerging trend of e-services opens up opportunities and challenges for end-users and organizations. Privacy and security are some of the challenges. The viability of e-services can be threatened if these are not correctly managed. This implies that the end-user might be concerned about the fact that an unknown third party can have access to the information that the user provides about himself/herself (Rust & Kannan, 2002). Despite the privacy and security concern, e-services can make it convenient for customers to access the service without constraints such as distance and opening hours (Rowley, 2006). 2.4.2 Website Navigation A challenging aspect faced by website designers is to reduce the disorientation when trying to locate desired information (Pilgrim, 2012). The provided tools from web browsers are limited and do not give information about the relationship between pages (Pilgrim, 2012). Nor does it help users to answer questions such as ’Where am I?’, ’Where can I go from here’ or ’Which pages points to this page’ (Bieber et al., 1997). 10 2. Background Farkas and Farkas (2000) studied the different guidelines related to the most com- mon navigational issues and synthesized these into their guidelines. The authors’ research is drawn from limited empirical research; however, it includes research and theory from human-computer interaction (HCI), interface design, cognitive psychol- ogy, and information design. Pilgrim (2012) investigated the frequently used website navigation tools from 2002 to 2011 and found three other dominating navigational support. Those were site search tools, sitemaps and index tools. Despite the dominating tools to provide effective navigation in websites, the author draws attention to the lack of detailed guidelines for designing these tools. Pilgrim (2012) states that disorientation within websites may never be solved. However, it can be reduced with the help of tools that aims to minimize the cognitive load when navigating. Pilgrim (2012), further states that future research needs to be done in this domain to help developers in their selection of website usability guidelines. 2.4.3 Accessibility Accessibility means that an interactive product can be used by as many people as possible (Sharp et al., 2019) and addresses that the user experience should be equivalent for people with and without disabilities (Web Accessibility Initiative, 2010). Web accessibility refers to the fact that people with disabilities should have the same experience of perceiving, understanding, navigating and interacting with websites (Web Accessibility Initiative, 2010). Moreover is the importance of incorporating accessibility from the beginning of a project to avoid redoing work (Lawton Henry, 2019). Some of the tools that increase the knowledge of web accessibility (Freire et al., 2007) are web browser plugins such as Funkify (n.d.), Voyant (n.d.), Webaim WebAim (n.d.) or digital products such as Microsoft Design toolkit (Microsoft Design, n.d.) and Cambridge Inclusive Design Toolkit (Clarkson & Coleman, 2007). Furthermore, a common framework and standard are the Web Content Accessibility Guidelines (WCAG) which has been approved as the international ISO-standard for web accessibility (Web Accessibility Initiative, 2010). The WCAG guidelines are grouped into: perceivable, operable, understandable, robust, and conformance. Each guideline has three levels of success criteria: A, AA and AAA. The latest update is the third version, WCAG 2.2, which is a working draft from the 13th of May 2021 that contains all the three levels of success criteria (Adams et al., 2021). Starting from 1st of January 2019, a new law about Web Accessibility Directive, (Lag (2018:1937) om tillgänglighet till digital offentlig service (Infrastrukturdeparte- mentet RSED DF, 2018)), also referred to ’DOS’-lagen, states that every public sector in Sweden has to follow the accessibility standards described in the European standard EN301549 (CEN CENELEC ETSI, 2015) which is based on the interna- tional standard WCAG level AA. The public sector in Sweden refers to authorities, county councils, municipalities, and other actors classified as bodies governed by 11 2. Background public law (DIGG, n.d.). The law applies to: • Web sites (including intranet and e-services) • Applications • Documents distributed from these platforms • Content that is spread through a third party (i.e. social media) The law also states that an accessibility report (’tillgänglighetsredogörelse’ in Swedish) has to be created by the public sector, which should describe how the current website or application follows the accessibility requirements (DIGG, n.d.). The Agency for Digital Government (DIGG) works to improve digital accessibility on websites and digital platforms controlled by the public sector (DIGG - Myndigheten för digital förvaltning, n.d.). 2.4.4 The Thesis’s Contribution The work presented in this thesis builds on earlier research to explore how to improve web navigation. Furthermore, existing guidelines in accessibility will be considered to create an accessible and usable product. While previous work focuses on general navigation to different information pages, and accessibility in general, this thesis will focus on a case study to further explore how to improve the navigation to certain e-services within a website. The main contribution of the thesis will help authorities and organizations gain more knowledge about how to design to help all users locate the correct e-services. 12 3 Theory The chapter covers relevant theoretical frameworks, concepts and related research to the thesis. It starts with an introduction about website navigation which aims to bring knowledge about navigation models, guidelines and navigation behaviors. The chapter follows up with theories about decision making to understand how humans make decisions in different situations. The theory about design systems is presented and describes the different terminologies used in the domain. Moreover, the chapter highlights the inclusive design methodology and covers definitions about disability, accessibility, and an introduction to the web accessibility directive. 3.1 Website Navigation Navigation plays an important role, not only for finding information on the web, but also to enhance understanding, reflect on the brand or organization and give credibility to the site (Kalbach, 2007). Website navigation can be defined in three ways (Kalbach, 2007, p. 5): 1. The theory and practice of how people move from page to page on the Web. 2. The process of goal-directed seeking and locating hyperlinked information; browsing the Web. 3. All of the links, labels, and other elements provide access to pages and help people orient themselves while interacting with a given website. Web navigation is about linking and creating relationships between different pages. Together, the elements form a system of navigation that help people find the infor- mation they are looking for. 3.1.1 Navigational Models Access to information on the web can take shape as different navigational models. Kalbach (2007) provides examples of some models to understand how information websites can be accessed. However, the author points out that nowadays, these models rarely exist alone but are somewhat mixed. 13 3. Theory ’The content-linking-only model’ describes how pages are linked to each other with- out any hierarchical organization or linking patterns (figure 3.1). In general, there is no traditional home page, and the links are embedded in the text. Although it may provide a strong relationship between pages, it does not provide high findabil- ity. This model will make it difficult to orient and search for information (Kalbach, 2007). Figure 3.1: The navigational model referred as the content-linking-only model. The Researchers’ own version of the original figure presented in Kalbach (2007, p. 6) ’The liquid information model’ characterizes a page where every word is an interac- tive element (figure 3.2). In other words, that means that all text are linked to other content pages. For instance, when clicking a word, an option menu can appear and refer to actions such as conducting a search, define the term or translate it (Kalbach, 2007). Figure 3.2: The navigational model referred as the liquid information model. The Researchers’ own version of the original figure presented in Kalbach (2007, p. 6) ’The filter model’ lets the user access all content through a single web page by filtering different chunks of materials (figure 3.3). The page provides interactive elements to let the user access the full content when clicking on it. All the content is presented on the same page, which means that the user never leaves the page, but instead keeps updating its content (Kalbach, 2007). 14 3. Theory Figure 3.3: The navigational model referred as the filter model. The Researchers’ own version of the original figure presented in Kalbach (2007, p. 7). ’The search model’ does not provide direct navigation or linking patterns to the internal pages (figure 3.4). To access the internal pages, the users have to go through three steps: (1) interact with a search system, and type in the correct keywords, (2) find the internal page in the presented search list and (3) click on the correct page. This model assumes that the item is known in advance and that the users know which keyword is practical to use (Kalbach, 2007). Figure 3.4: The navigational model referred as the search model. The Researchers’ own version of the original figure presented in Kalbach (2007, p. 8) ’The structural browse model’ takes form as a hierarchical tree structure where a set of links are visually represented and separated from the page (figure 3.5). These links provide branches to the tree, which provides access to more information (Kalbach, 2007). Figure 3.5: The navigational model referred as the structural browse. The Re- searchers’ own version of the original figure presented in Kalbach (2007, p. 9) 15 3. Theory 3.1.2 Website Navigation - Design Guidelines Design guidelines are useful as a framework for designers to make grounded decisions in situations where there are limitations in resources hindering an examination of every feature (Pilgrim, 2012). According to Pilgrim (2012) there are mainly two types of guidelines; detailed low-level rules and high-level guiding principles. Com- monly, different guidelines become too general or too narrow, resulting in difficulties implementing them in specific projects. There are limited guidelines focused on the navigational aspects of the websites (Pilgrim, 2012). However, some of the general guidelines of web design have more focus on guiding the user towards finding the relevant information. Horton et al. (2016) describes the following; • The terminology and organization structure of websites should match the users mental model to help them navigate the site. Websites that order their content according to how the management works often confuse the user and make it hard to find the content. • When guiding the user to find what they are looking for, it is good practice to segment the information into short chunks. Bulleted lists, tables, titles and subtitles together with short information enables quick scanning. • The user expect hyperlinks to lead them directly to related information. If this need should be met, the chunks need to be structured and organized placing related information together. A logical consistent site organization allow the user to predict where to find things. • The design should be made with ’enough’ levels in the menus since too many branches buries the information deep and requires too many clicks. Further- more, users tend to be confused by deep hierarchies. • For sites bigger than a few dozen pages, the users expect a functional search function. Browsing can never assure the user that they found all relevant information about a subject indicating the importance of the search box. Pilgrim (2012) indicate that the web navigational guidelines are deficient, lacking advice of specific navigation. 3.1.3 Information Foraging A theory of how people navigate on the web and how they behave can be explained with Information Foraging, developed by Peter Pirolli and Stuart Card (Budiu, 2019). The theory was inspired by animal behavior, more specifically, how they forge for food which shares a similar language with information foraging as illustrated in table 3.1. The theory of information foraging states that when users have a goal to find information, they estimate and choose the candidates they believe can maximize the rate of gain (Pirolli & Card, 1999): 16 3. Theory Rate of gain = Information V alue Cost associated with obtaining that information In general, the equation means that people will make decisions based on (1) the likeliness that the page will lead them to their goal and (2) the time it takes to get to the goal. In other words, it means that the theory can explain why people do not visit every single link or keeps scrolling: they want to spend as little time possible to get as much information they can, thus maximize the Rate of gain (Budiu, 2019). Animal Foraging Food A site containing one or more potential sources of food. Search for food The animal’s assessment of how likely the patch will provide food The totality of food types that an animal may consider in order to satisfy hunger Information A website (or other source of information) Search for information How promising a potential source of information appears to the user The totality of the information sources that a user may consider to satisfy information need Information Foraging Goal Diet Scent Forage Patch Table 3.1: The common language between animal foraging and information for- aging. The Researchers’ own version of the original figure presented in (Budiu, 2019). Information Scent One of the factors affecting the Rate of gain equation is the ’information scent’, which was popularized by Spool et al. (2004). The concept of information scent plays an important role when users search for information on the web. The user estimates how well the web page suits their information need by judging the ’scent’ of how likely it is to find the information. The scent of a web page is the titles, images, and information displayed on the site. For instance, a user that is looking for dish towels but lands on a page displaying images of candies can, by judging the information scent understand that information 17 3. Theory about towels will not be found there (Budiu, 2019). Spool et al. (2004) means that people usually scan for their trigger words (words or phrases similar to what they are seeking) which can give them a sense of the scent. The Cost-Benefit Perspective Another factor affecting the rate of gain equation is the cost associated with obtain- ing that information (Budiu, 2019), which can be explained with the cost-benefit perspective (Katz & Byrne, 2003). In general, the cost-benefit perspective describes the human decision-making process. It consists of two factors: the cognitive effort to use a strategy and the rate of accuracy in the results it can provide (Kleinmuntz & Schkade, 1993). For the information foraging, that means the cost of time and effort involved to extract the information and the cost of opportunity. The cost of opportunity relates to the fact that once something is decided, the potential oppor- tunity to look for others will be lost (Budiu, 2019). The cost of time and effort involves the two types of user activities (Budiu, 2019): (1) Between-patch activities: Finding and gathering sources (patches) of information. (2) Within-patch activities: Inspecting the sources (patches) to gain information. Enrichments The two activities describing the cost of time in the cost-benefit perspective (section 3.1.3) can be affected by adaptations called enrichments. That refers to a type of interaction, behavior or strategy to maximize the information foraging. In other words, enrichments can be described as a tool that is used for foraging information which can either be a learned behavior (behavioral enrichments) or created to the specific patch where the user has to invest time and build it (interaction enrichments) (Budiu, 2019). ’Behavior enrichments’ are learned tools that have proved to be successful in helping the user extract information efficiently. An example is the ’F-pattern scanning’ where users can search for a specific word or sentence on the web page and effectively scan the information scent without reading the entire page. Pernice (2017) lists more scanning patterns which are summarised in table 3.2 18 3. Theory Scanning Pattern Layer-cake pattern Spotted pattern Marking pattern Bypassing pattern Commitment pattern The eyes focus on the headings and subheadings but skips the normal text below. If the behaviour was plotted on a heat map, it would show horizontal lines resembeling a cake. User scans the information and skips big chunks of text to look for specific elements such as links, digits or a word. The eyes are focused in one place while scrolling or swiping the page. It happens more often on a moblie than on desktop. When user skips the first word of a text where the list starts with the same words. When the user reads through the whole page such as all paragraphs and contents. Description Table 3.2: Behavior enrichments can be described with different types of scanning patterns. The Researchers’ own table. In turn, ’Interaction enrichments’ refers to the amount of effort a user has to find information effectively. For instance, the time users spend thinking of keywords that can increase the relevant searches (Budiu, 2019). Pirolli and Card (1999) highlights the importance of considering designing websites that can maximize the efficiency of information foraging without the use of enrich- ments. Budiu (2019) provides examples that optimize the websites, such as using systems that rank search results and put the relevant results first. Other optimiza- tion tools are auto-suggestions (Moran, 2018), and scanning-friendly formatting such as descriptive headlines, italic or bolded words and bulleted lists (Budiu, 2019). 19 3. Theory The Attentional Capture Perspective The attentional capture perspective can become a critical factor in the choices users make on the web. It refers to how visual elements in the environment can draw (or ’capture’) attention and therefore affect the decision making (Todd & Kramer, 1993). It could be colors, brightness, and objects on or close to visual boundaries. A feature that catches a user’s attention will immediately signal that it is a ’landmark’ on the page. For instance, if a search bar draws attention with its visual cues, the user may perceive that the search bar is the primary way to find items and that other features are secondary alternatives (Katz & Byrne, 2003). In addition to that, Burke et al. (2005) studied how commercial banner ads affected the search time on a web page and concluded that it hindered the search even though the users did not look at the ads. 3.2 Decision Making Theory Information is all around us, everywhere. To handle this information overload, hu- mans need to filter the information and the stimuli to a reasonable amount (Stern- berg et al., 2011). Decision making works the same way - by reducing the available information, the human makes it possible to make decisions. When navigating a website, many decisions need to be made about where to click, what to search for, and where to direct the focus. Heuristics is an aid in these circumstances, allowing the brain to examine fewer cues and simplify the design making process (Sternberg et al., 2011). Generally, a heuristic could be seen as a rule of thumb, helping the person use previous experiences to act more quickly. However, beyond reducing the cognitive load, the shortcuts also open up for biases and errors since much informa- tion falls out (Caraban et al., 2019; Sternberg et al., 2011). This section will present some heuristics and possible biases. 3.2.1 Heuristics There are several heuristics used in different situations. ’Satisfacing’ is one of them, relating to the idea that humans get as close as possible to optimizing a decision without really optimizing (Sternberg et al., 2011). The idea is that because of lim- ited resources and time, and it is too expensive to make the best decision. However, the decision is optimized to the resources the person has. When using the satisfic- ing heuristic, the person selects the first option that meets the minimum level of acceptability, i.e. it is ’good enough’ (Sternberg et al., 2011). Other heuristics are ’Elimination by Aspects’ and ’Framing’ (Sternberg et al., 2011). Elimination by Aspects treats the problem of too many options that cannot be considered in the available time. By focusing on one aspect at a time, e.g. "needs to be cheaper than 500 SEK", the person can efficiently eliminate many options and end up with a reasonable amount. The Framing Heuristic implicate that how different options are represented affect how they are selected (Sternberg et al., 2011). For example, humans tend to prefer certain gains rather than uncertainty. If a problem 20 3. Theory is stated and formulated so that one option seems more safe and more certain than the others, that option is more likely to be selected. In user interface design, Jakob Nielsen has formed ten usability heuristics that have acted as rules of thumb for human-computer interaction (Nielsen, 2020). An example is the heuristic ’Error Prevention’ that prevents errors and can be done by supporting undo and giving warnings to users. A similar heuristic is ’Help users recognize, diagnose, and recover from errors’, which means that error messages should be expressed in plain language where the problem is stated, and a solution is suggested. All usability heuristics are presented in 3.6. 10 Usability Heuristics By Jakob Nielsen #1 Visibility of System Status #2 Match Between System And the Real World #3 User Control And Freedom #4 Consistency And Standards #5 Error Prevention #6 Recognition Rather Than Recall #7 Flexibility and Efficiency of Use #8 Aesthetic and Minimalist Design #9 Help Users Recognize, Diagnose, And Recover From Errors #10 Help and Documentation Figure 3.6: The 10 usability heuristics by Jakob Nielsen presented in Nielsen (2020). The Researchers’ own illustration. 3.2.2 Biases As mentioned, decisions made by heuristics might lead to unintended biases (Cara- ban et al., 2019; Sternberg et al., 2011). Some of these biases are ’Overconfidence’, ’Illusory Correlation’ and ’Hindsight Bias’. Research has shown that people tend to be overconfident, either of their skills, judg- ment or knowledge (Sternberg et al., 2011). If a poor decision is made because of this overconfidence, it is called the overconfidence bias. Exactly why people tend to be overconfident is unclear. However, it might be because people are unaware of how little they know or that their resource is unreliable (Sternberg et al., 2011). People tend to create correlations and relationships between things that do not correlate, so-called illusory correlation (Sternberg et al., 2011). It could be events, categories or attributes. For example, people might believe that a specific cause affected some- 21 3. Theory thing while it was unconnected. The bias might also impact the thought of other people and form stereotypes, making it easier to prejudge. When looking at a happening in retrospect, it is easy to believe that the outcome could have been figured out beforehand while noticing all the signs (Sternberg et al., 2011). However, research indicates that humans rarely recognize the signs while they are happening; thus, first afterwards feel that the outcome was obvious. This phenomenon is called the hindsight bias (Sternberg et al., 2011). 3.3 Design System The design system does not have a standard definition, and people within the web community use it differently (Kholmatova, 2017). According to Kholmatova (2017, page 25), a design system can be defined as "a set of interconnected patterns and shared practices coherently organized to serve the purpose of a digital product". In other terms, it is a framework consisting of style guides and pattern library (Fitzpatrick, n.d.) as seen in figure 3.7. A style guide is a part of the design system and defines the rules of branding and visual style of a product (Mockplus, 2020). It consists of guidelines for typography, icons, color, etcetera (Mockplus, 2020). A pattern library is a collection of recurring and reusable patterns that are used to solve design problems (Kholmatova, 2017). It consists of several building blocks (Fitzpatrick, n.d.). One of them is ’Identity’, which aligns with the style guide such as logos, fonts and colors. ’Components’ help define the user flows and take form as a login structure, survey form, or error screen. ’Elements’ are smaller parts of the components such as buttons, form fields and dropdowns. Elements can use ’Gestalt Principles of Visual Perception’ to describe how a specific arrangement of content can help people determine if certain elements belong to the same group. An example is the principle of proximity, which explain how placing elements near each other can provide visual cues that these are grouped (Harley, 2020). Another block is ’Interactions’ that shows the state of an activity, for instance, hover states, page loads, or slide effects. One way of showing interactions is to use ’Affordances’ which provide visual cues on how to interact with an object. All these patterns create a design language of the interface and can be viewed as a vocabulary of design elements in terms of shape, color, typography and the combination of these (Cooper et al., 2014). 22 3. Theory Style Guides Design System Pattern Library Identity Components Elements Interactions Figure 3.7: Overview of a design system that can include both a style guide and a pattern library. The Researchers’ own figure. 3.4 Usability According to the ISO-standard of Usability (ISO 9241-210:2019, 2019, p. 3), usability is defined as the "extent to which a system, product or service can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use", and goes in line with the five components of usability presented by Nielsen (Nielsen, 2012). These are learnability, efficiency, memorability, errors, and satisfaction. Several methods and approaches can improve usability. The fundamental approach is usability testing, which consists of three modules (figure 3.8). One of the modules is to find representative users who know about the user group’s needs, opinions, struggles, and preferences. The following module is to let the users perform repre- sentative tasks and provide context to engage with the interface. The third module involves observing the users and identify what they do, where they succeed and what difficulties appear when interacting with it (Nielsen, 2012). Find representative users Find representative tasks Observe users The Three Modules of Usability Testing 1 32 Figure 3.8: The three modules of usability testing. The Researchers’ own figure. 23 3. Theory Nielsen and Landauer (1993) have researched the ultimate amount of users involved in usability testing. Their conclusion is presented with a graph (figure 3.9) that illustrates how the amount of users corresponds with the percentage of identified usability problems. In their final analysis, Nielsen and Landauer (1993) argues that five users are enough to capture 85 percent of the usability issues. The equation acknowledges that involving more than five users will fall into the risk of not finding or learning anything new and instead becoming costly. However, the best scenario is to run as many small tests as possible within the budget and the time frame. Numbers of Test Users 0 0% 25% 50% 75% 100% 3 6 9 12 15 U sa b ili ty P ro b le m s F o u n d Figure 3.9: The usability graph that illustrates how the amount of test users corresponds to the percentage of identified usability problems. The Researchers’ own version of the graph and original data presented in Nielsen and Landauer (1993, p. 209). 3.5 Inclusive Design Inclusive design is a methodology that strives to include and learn from people with different perspectives by encouraging the creation of products, systems and services in a diversity of ways for people with and without impairments. The most important part is to make it possible for everyone to participate with a feeling of inclusion and belonging (University of Cambridge, n.d.). Moreover, Microsoft Design (n.d.) has created a table (figure 3.10) that divides the different dimensions and impairments into touch, see, hear and speak with an example of the respective time dimensions: permanent, temporary and situational. The purpose of the table is to illustrate the range of different impairments that can lead to disability. 24 3. Theory Figure 3.10: The persona spectrum by Microsoft Design (n.d.) under CC BY-NC- ND. There are principles aimed to make it easy for everyone to have an inclusive design approach. The principles are found in Microsoft’s inclusive design toolkit (Microsoft Design, n.d.) and present the following: • Recognize exclusion: exclusion happens due to mismatched interactions be- tween a person and society. By identifying the factors that lead to exclusion, designers can generate ideas and solutions for interactions between a wider range of people. • Learn from diversity: when products, systems or services do not perform well for people with a set of disabilities, these people adapt to the situation. By learning from diverse people, insights will be gathered to understand how people adapt to new situations and the emotional context around them. • Solve for one, extend to many: exclusion can be experienced by all people, anywhere, anytime. The principle highlights the notion of the persona spec- trum. It emphasizes that a solution for one person with a permanent disability will likely become a solution for the people with temporary and/or situational disability. 3.5.1 Disability Disability can be defined with the medical model and the social model. The medical model claims that people are disabled because of their impairment, such as visual, 25 3. Theory cognitive, hearing, or physical impairment. On the contrary, the social model states that people are disabled because of society and the surroundings. In recent years, the latter model has been widely used to define the term disability, and World Health Organization (2020) describes it as "an umbrella term for impairments, activity limitations, and participation restrictions". Siebers (2008) writes that the issue is created by environmental and social barriers rather than the impairment of a person. A product, system or service created in a way that necessitates a specific type of interaction can make it unusable for someone with a disability, thus leading to a new disability. In this context, disabilities were created due to the poor design of the interaction between the user and the technology, not due to the disability itself (Sharp et al., 2019). In Sweden, there are no statistics on the number of people who face society’s chal- lenges due to their disability. There are different types of impairments, and the Swedish government is not allowed to collect information on various disabilities among the population. Furthermore, some disabilities are hard to diagnose while others only affect the person during some periods of life (Funka, n.d.). However, World Health Organization (2020) estimates that over 1 billion people worldwide live with some form of disability which equals about 15 percent of the world’s pop- ulation. According to Sharp et al. (2019), the number of people born with some disability is fewer than 20 percent. At the same time, it is estimated that 80 percent of people will have any disability when they reach 85. 26 4 Methodology This chapter provides a description of relevant frameworks and methods for the project. The beginning of the chapter introduces the concept of research through design to understand why design activities are helpful to solve problems. It continues with a section on remote research, user research methods, and the design process intended to be used in the thesis. Furthermore, the chapter introduces a list of different methods considered helpful for answering the research questions. 4.1 Research Through Design Design is a broad definition, which might refer to both the design process and the final design of a product or prototype. The combination of ’design’ and ’research’ in human-computer interaction has been developed for many years (Zimmerman et al., 2007). It has previously been regarded as separate areas, where design comes from the industrial practice and research from academic experiences (Stappers & Giaccardi, n.d.). Research through design can, at first sight, seem non-academic be- cause of its exploratory manner. However, it is also possible to see it from another perspective and worship the design research with its aptitude for exploration and speculation (Gaver, 2012). Gaver (2012) explains that every design-related research is unique and often covers statements such as ’what might be’ instead of ’what is’. Accordingly, the design problems can often not rely on doing traditional scientific hypotheses testing. The goal is instead to reach the point where theories are ’some- times right’. Stappers and Giaccardi (n.d.) means that research through design can be viewed as design activities that contribute to new knowledge. By using the design process, practitioners can understand, frame and work in iterations to develop pro- totypes that address the problem. The notion of research through design is related to Rittel and Webber (1973)’s definition of wicked problems. 4.1.1 Wicked Problems Rittel and Webber (1973) categorize problems into two areas: tame and wicked problems. The definition of tame problems is the ones that can be clearly defined and solved by applying the correct methods. In contrast, wicked problems are more complex and cannot be properly formulated (Rittel & Webber, 1973). Furthermore, Churchman (1967) states that wicked problems are a: 27 4. Methodology " (...) class of social system problems which are ill-formulated, where the information is confusing, where there are many clients and decision- makers with conflicting values, and where the ramifications in the whole system are thoroughly confusing." (Churchman, 1967, p. 141) In other words, a wicked problem consists of many variables that would be impossible to control and therefore seem impossible to solve (Interaction Design Foundation, n.d.-b). Rittel and Webber (1973) explains wicked problems by taking a freeway as an ex- ample. It is expressed that it would not be feasible to start building a freeway and then later change things after its performance. Nevertheless, it is not feasible to test the freeway in all scenarios. Somewhere in that process, a decision has to be made. Designers are often dealing with wicked problems, and one way to solve them is to use Design thinking since it can be re-framed into a human-centered approach and make it easier to prioritize the users’ needs (Interaction Design Foundation, n.d.-a). 4.2 Remote Research The project was written during the spring of 2021 when Covid-19 was widely spread worldwide, inevitably forcing everyone who had the opportunity to work from home. Due to these circumstances, the researchers focused on methods that could be done digitally - remote research. Remote research means researching without interacting in-person or face-to-face (Asjes, 2014). The research is often performed using tools such as a computer or phone, allowing both the researcher and participant to see and hear each other. Some software provides screen-sharing, which makes it possible to see what the other person is seeing and interacting with (Asjes, 2014). Preece et al. (2015) highlight the advantages and disadvantages of remote research and state different aspects to consider before starting. Poor connections can cause difficulties in video and acoustics performance, including a risk that the participant might multitask instead of focusing on the session. Moreover is the aspect of security, where sensitive information or confidential materials can be a concern when doing research remotely (Tulathimutte & Bolt, 2010). Another concern might be if special equipment such as software or physical equipment are required to run the study (Tulathimutte & Bolt, 2010). Likewise is the aspect of having to install and configure tools, which can become complicated for first-time users or people that have lower computer literacy (Asjes, 2014). Despite the disadvantages, remote research also opens up advantages and opportunities. There is no need to travel (Preece et al., 2015) which can lead to geographic diversity (Tulathimutte & Bolt, 2010), and participants are in their habitat, which can make them feel more relaxed (Preece et al., 2015). In addition, participants can leave the conversation, which can make them feel more controlled and secure (Preece et al., 2015). 28 4. Methodology 4.3 User Research Methods Interaction design includes different methods of user research. The methods range from data gathering to data analysis and can be classified into different dimensions such as: qualitative and quantitative, or behavioral and attitudinal. A coverage of these methods is essential in the design process, and one way to break down the methods is to use a chart suggested by Rohrer (2014). The chart consists of two axes defining both Qualitative vs. Quantitative and Behavioral vs. Attitudinal. An example is given with some of the methods that are considered for the project where the methods about the dimensions are referred from Martin and Hanington (2012), seen in figure 4.1. Even if the chart gives an overview of the methods’ positions in relation to the dimensions, Preece et al. (2015) argues the importance of understanding that all types of data gathering methods can result in both qualitative and quantitative data. What people do Why and how to fix BEHAVIORAL ATTITUDINAL QUALITATIVE (DIRECT) QUANTITATIVE (INDIRECT) How many and how much What people say • Affinity diagramming • Persona • Scenario • Digital prototyping (hi-fi) • Simulation • Think aloud • Cognitive walkthrough • A/B Testing • Questionnaires • Interview • Questionnaires • Questionnaires • Eye-tracking • Questionnaires Figure 4.1: A method matrix with the dimensions of qualitative, quantitative, attitudinal and behavioral aspects. Authors’ own version of the figure inspired by Rohrer (2014). Quantitative research is well suited to understand problems related to ’how many’ and ’how much’ (Rohrer, 2014). The data often includes numbers and is based on the performance of a task, such as success rate or a number of errors. The context 29 4. Methodology of the problem might be lost since the result often is presented as numbers, making it harder to interpret it without a reference point (Budiu, 2017). The findings often need to complement qualitative research, which answers questions related to ’why’ or ’how to fix a problem’. It might be data with detailed verbal descriptions (Baxter et al., 2015), where the outcome is a set of words and images (Preece et al., 2015). The attitudinal dimension deals with what people say (Rohrer, 2014) and serves the purpose of understanding peoples’ opinions, feelings, beliefs, or thoughts (MBA Skool Team, 2020). In other words, what people express. The contrast of attitudinal is behavioral, which gives insights into what people do (Rohrer, 2014). Scott (2017) highlights the importance of combining attitudinal and behavioral data since they supplement each other. As an example, it is common for participants to behave in ways that are different from their expressions (Scott, 2017). 4.4 Design Thinking It is important to understand and address changes in the users’ behaviors, and the environment around them (Interaction Design Foundation, n.d.-a). Interaction De- sign Foundation (n.d.-a) mention that wicked problems (section 4.1.1) are increasing and that design thinking is one way to target the problem; by breaking down the problem in smaller pieces and focusing on the user. It could be seen as a five-step process, often performed in iterations or parallel to each other (Interaction Design Foundation, n.d.-a). The five steps are Empathize, Define, Ideate, Prototype & Test (Figure 4.2) (Siang & Dam, 2021). Empathize Define Ideate Prototype Test Figure 4.2: The five stages of the design thinking process which can include iter- ations of phases. Authors’ own figure. 4.4.1 Empathize Empathize is the first stage of the design thinking process to capture an empathetic understanding of the problem. This can be done by engaging and immersing with the users to understand their feelings, experiences, and motivations (Siang & Dam, 2021). The goal is to discover users’ explicit and implicit needs and find the gaps between what users do and what they say they do (Siang & Dam, 2020). Siang and Dam (2021) highlights the importance of applying empathy in the human-centered design process to allow designers and researchers to leave assumptions about the world behind and instead focus on other peoples’ actual needs. Many methods can be used in the empathize phase. This section will cover the methods illustrated in figure 4.3. 30 4. Methodology Simulation Exercises Indirect Observation Eye Tracking InterviewThink Aloud Cognitive Walkthrough QuestionnaireLiterature Reviews Empathize Figure 4.3: The eight methods involved in the empathize stage. Authors’ own figure. Literature Reviews Literature reviews is a research method and a fundamental part of academic papers where previous research enables contextualization and foundation of the designated area (Snyder, 2019). The method is a practical approach to provide answers, give an overview of the research area, identify gaps, discuss a specific topic, or laying the foundation to a new conceptual model or theory (Snyder, 2019). Furthermore, Martin and Hanington (2012) writes that literature reviews can be freestanding or a component of a research paper. Simulation Exercises Simulation exercises are established in multiple professions, offering circumstances that are hard to create otherwise (Martin & Hanington, 2012). According to Martin and Hanington (2012), the primary purpose of the simulations is to evoke user empathy and affect the design decisions through experience. The simulations can be created and designed in different ways, mainly treating physical disabilities and limitations, e.g., cognition deficits, brain injuries, and visual impairments (Martin & Hanington, 2012). However, a limitation is that an experience from a simulation exercise might not always correspond to reality. Think Aloud Think aloud is a method used for evaluating existing products, prototypes, ser- vices, or systems. The participant is asked to articulate their thoughts, feelings, and actions while simultaneously performing some task (Martin & Hanington, 2012). While taking part in the participants’ actions and reasoning, the designer gets valu- able data about struggles and frustrations which enables improvements and correc- tion (Martin & Hanington, 2012). 31 4. Methodology Eye Tracking Eye tracking is a method measuring technical information about where, when, and for how long a participant is looking at something (Martin & Hanington, 2012). According to Martin and Hanington (2012) it helps examine participant’s naviga- tion on websites. The eye movements are identified as fixations or saccades (rapid movements between points), and a heat map is created visualizing the result. The heat map does not explain the underlying motivations of the user, thus making it suitable to perform together with complementary methods (Martin & Hanington, 2012). Interview Interviews can be divided into four categories: open-ended/unstructured, semi- structured, structured, and group interviews (Sharp et al., 2019). If the purpose is to get a first impression about something, an informal and open-ended interview is often suitable (Sharp et al., 2019). Even though an unstructured interview is open for new perspectives and questions, it is recommended to prepare an agenda and discuss topics. It is important to remember that the data will not be coherent be- tween different participants since the interviews develop their format (Sharp et al., 2019). This might make it harder to analyze. Interviews conducted remotely are increasing in popularity. It might be ensuring for the participant to be in their environment without worrying about the circumstances around them, such as other participants, clothing, and the possibility of leaving the conversation (Sharp et al., 2019). Indirect Observation The indirect observation could be performed in different ways, either by analyzing textual material from observations or directly from narratives such as tweets and forum posts (Anguera et al., 2018). According to Anguera et al. (2018) it can be helpful to mix non-spontaneous behavior, such as questionnaire responses, with data from more natural circumstances to get value from both the qualitative and quantitative data. Indirect observation in terms of examining trails of activity and scraping data is an unobtrusive method. However, it might raise ethical questions if the user is unaware of the activity (Sharp et al., 2019). Cognitive Walkthrough One way to examine prototypes is to let a usability expert perform a cognitive walk- through from the user’s point of view. The method focuses on user interaction by examining how a new user can perform specific tasks in a system (Interaction Design Foundation, 2020). The first step is to define the tasks that should be evaluated and divide them into smaller subtasks (i.e., ’open the browser’). Preferably, one can ask four questions after each subtask; "Will the user try and achieve the right 32 4. Methodology outcome?", "Will the user notice that the correct action is available to them?", "Will the user associate the correct action with the outcome they expect to achieve?" and "If the correct action is performed; will the user see that progress is being made towards their intended outcome?"(Interaction Design Foundation, 2020). Questionnaires Questionnaires are simple to produce and the dominant tool to collect survey in- formation (Gillham, 2008). Martin and Hanington (2012) states that it could be used independently and in integration with other methods to supplement the data. The authors further highlight the importance of considering how the questionnaire is structured. Open-ended questions enable depth of response, while closed-ended questions provide an opportunity for numerical analysis. Likert scale questions are highly recommended to keep both neutrality in question but also receiving strength of response. In general, the method can be used in various phases for different pur- poses (Martin & Hanington, 2012). A disadvantage with the method is that it does not give any space for asking follow up questions directly to the user. 4.4.2 Define The second stage is used to gather the data from stage one and analyze it (Inter- action Design Foundation, n.d.-a). The defined core problems are called problem statements and should be human-centered (Siang & Dam, 2021). Furthermore, af- ter defining the problems, the designers might accelerate against the third stage by introducing questions as "How might we...?" (Siang & Dam, 2021). The section will cover the methods illustrated in figure 4.4. Persona Behavior Archetypes Affinity Diagramming Design Requirements Scenario Define Figure 4.4: The six methods involved in the define stage. Authors’ own figure. Affinity Diagramming Affinity diagramming is used to organize and structure the data from earlier research (Martin & Hanington, 2012). Researched-based insights are noted on separate sticky notes, which later are clustered according to their similarities. Affinity diagramming is an inductive method, creating clusters and groups gradually instead of using predefined categories (Martin & Hanington, 2012). 33 4. Methodology Persona Personas help present the goals and needs that a product or service should be de- signed against (Preece et al., 2015). They can be used to communicate and express the purpose and vision (Sharp et al., 2019). Personas do not describe real users; instead, they are meant to be realistic representations of the typical user. Behind each persona, there is a synthesis of real users. By including detailed information, such as name, personal details, and photographs, the persona feels like a real person - making it easier for the designer to see them as potential users (Sharp et al., 2019). This method supports inclusive design where accessibility personas can be created to establish an understanding of people with disabilities. A risk with the method is that the demographical data can lead to biases and assumptions, which should be avoided. Scenario Scenarios could be described as a combination of things that might happen (Good- win, 2009). They can be used for different purposes, including creating requirements and checking interaction options. Goal-Directed scenarios are making use of the per- sonas by describing their future interaction with a product, thus giving them value and explains their motivations for a particular behavior (Goodwin, 2009). When scenarios are used to define requirements, called context scenarios, it is still early in the process. Thus the focus naturally lies on high-level functions, portraying the ideal system behavior for different situations. Behavioral Archetype Behavioral archetype is a method similar to Persona (section 4.4.2) with the differ- ence that, instead of presenting a person based on demographics such as age and gender, behavioral archetypes represent and present the behavior and motivation of the users. This is done by capturing their motivations, goals and general attitudes (Doneva, 2017). A toolkit by Smashing Ideas (2017) explains the creation of behavioral archetypes which can start with a determination of the goals, needs, pain points, thoughts, feelings, and actions of the user is to the context or scenario they are active in. The method provides a foundation to prioritize features of a product so that it will match the users’ behaviors (Doneva, 2017). Design Requirements While defining a description of the service or product that should be designed and developed, requirement activities might be helpful to gather insights in terms of requirements. A requirement could be defined as a statement that specifies how a product should perform or what it should do (Sharp et al., 2019). The discovery of requirements is done continually and is an iterative process. Sharp et al. (2019) states that it is necessary to capture all iterations’ requirements to ensure that noth- ing gets lost in the process. A consequence of limited notations of the requirements 34 4. Methodology might be miscommunication and misunderstanding. Traditionally, it is possible to divide requirements into two categorizations - func- tional and nonfunctional. The functional requirements determine what the product shall do, while the nonfunctional express constraints or characteristics. However, more categorizations exist, and they vary between different contexts. 4.4.3 Ideate The focus of stage three is to ideate and generate ideas (Interaction Design Founda- tion, n.d.-a). The knowledge from previous steps is used to challenge assumptions and brainstorm potential solutions (Siang & Dam, 2021). Siang and Dam (2021) highlights the importance of examining multiple ideas and problem solutions while ’thinking outside the box’. The section will cover the methods illustrated in figure 4.5. Sitemap BrainstormingFlowchart Ideate Sketching Figure 4.5: The three methods involved in the ideate stage. Authors’ own figure. Sitemap Sitemaps can be used in the design process to help with the organization of content and the navigation (Babich, 2019). There are two types of sitemaps - visual ones representing the site organization and coded ones showing a list in XML. The visual sitemap often shows a hierarchical view of the relations between pages (and sub- pages) and helps to identify how items might be grouped. The purpose of this structured representation is to create a better information architecture. The production of a sitemap is done by creating blocks and connecting them with lines (Babich, 2019). The sitemap should include a home page and items. The items can have numbers correlating to their position in the hierarchy and a label that explains more about them. Flowchart Flowchart represents sequences of steps in a process (Chapin, 2003) that can show the whole process, the different steps, the relationship between the steps, as well as the beginning and end of a process. Flowcharts are a helpful tool to communicate 35 4. Methodology and create an understanding of a process. To build a flowchart, standard symbols are used to represent specific actions (Associates Staff, 1995). Brainstorming Brainstorming involves combining a group of people, creating possibilities and ideas (Goodwin, 2009). Even though many of the generated ideas are a bit crazy and perhaps built on faulty assumptions, it is good to write them down and get them out of the system. It is essential to have an open mind and not criticize or comment on the ideas too much since it will hinder creativity - the idea might develop into something suitable later on. A brainstorming session should be one to two hours at most; then, the interesting ideas should be noted and saved for later (Goodwin, 2009). Sketching Sketching is a method used for creating the ground of the design work (Arvola, 2014). According to Arvola (2014) a sketch could be seen as a suggestion, an exploratory action that asks questions. It is a cheap method, meaning that the investment is low, thus opening up for creating many versions and getting a better understanding of the future product (Buxton & Buxton, 2007). Cognition could be seen as an interactive process between the mind and the world, making it natural to use sketching as a supportive tool for the thinking process (Gedenryd, 1998). Gedenryd (1998) states that sketching most often is performed in a simple context with elementary tools, such as pen and paper. Drawings and sketches help the human mind to put things in context and see if ideas are doable (Gedenryd, 1998). In a design process, it is essential to create many sketches to expand the ideas and avoid selecting the first standard ideas (Arvola, 2014; Traynor, 2012). The sketches should be annotated to enable further development and examination (Arvola, 2014). 4.4.4 Prototype Stage four involves a more experimental standpoint to identify the most suitable solution for each problem (Interaction Design Foundation, n.d.-a). Inexpensive ver- sions of the product, or specific features of the product, should be produced to in- vestigate the solutions created in the previous stage (Siang & Dam, 2021). Without prototypes, the designers would produce the final product and not until afterward realize the demands of the product (Arvola, 2014). According to Arvola (2014), using prototypes during the process makes small changes easier to implement and test before developing the final product. The purpose of the stage is to highlight potential constraints and problems combined with better insights about how the users might behave and feel for the final product (Siang & Dam, 2021). The section will cover the methods illustrated in figure 4.6. 36 4. Methodology Digital Prototyping Paper Prototyping Prototype Figure 4.6: The two methods involved in the prototype stage. Authors’ own figure. Digital Prototyping (High-Fidelity Prototyping) A high-fidelity prototype often provides much functionality and involves the look and feel, making it similar to the final product (Sharp et al., 2019). This can be compared to low fidelity prototypes that often are made of paper or similar. It could be seen as a scale, where some prototypes might be of higher fidelity than another, even though both of them could be classified as low fidelity. The main idea is that the prototype needs to have enough fidelity to test the wanted functions. One way to handle it is to follow the economic principle: "The best prototype is one that, in the simplest and most efficient way, makes the possibilities and limitations of a design idea visible and mea- surable." (Lim et al., 2008, 7:3). There are both positive and negative aspects of high-fidelity. The positive part is that it offers almost complete functionality, the look and feel of the intended product, is fully interactive, and can easily be used for testing and marketing (Sharp et al., 2019). However, it is also more resource-heavy during development, time- consuming, and might be mistaken for the final product while testing and creating false expectations. High-fidelity prototypes do not need to be produced from scratch. Instead, it might be efficient to create them by combining existing components from libraries, developer kits, or similar (Sharp et al., 2019). Paper Prototyping (Low-Fidelity Prototyping) A low fidelity prototype is a simpler version of the final product and could, as an example, consist of paper or cardboard (Sharp et al., 2019). The simple and cheap production makes paper prototypes suitable for exploring different ideas (Sharp et al., 2019). 4.4.5 Test The final stage involves comprehensive testing of the prototype (Interaction Design Foundation, n.d.-a). The results from this stage are often used to redesign existing solutions from previous steps, working with the stages in iterations (Siang & Dam, 2021). The section will cover the methods illustrated in figure 4.7. 37 4. Methodology A/B Testing Think Aloud Eye TrackingCognitive Walkthrough Test A B Interview Figure 4.7: The five methods involved in the test stage. Authors’ own figure. A/B Testing A/B tests are used to put two different versions of the same product or prototype against each other and compare them in terms of which performs better in a deter- mined goal (Martin & Hanington, 2012). An A/B test often consists of a hypothesis, a controlled test, and an altered test. The hypothesis is the main question that is tested to confirm whether it is true or not (Stepanov, 2018). The controlled test is referred to ’Test A’ and will provide data to compare with the altered test. The altered test is referred to ’Test B’, which is where the changes are implemented. This setup will make it possible to conclude a hypothesis (Farmen, n.d.). Usability Testing As mentioned in the theory about ’usability’, Nielsen and Landauer (1993) means that usability testing consists of the three modules: (1) find representative users, (2) let the users perform representative tasks, and (3) observe the users. Different methods can be used to conduct usability testing. Some of the methods that can be used are described earlier in the empathize phase, such as a cognitive walkthrough, think aloud, eye tracking, and interviews. These methods can be reused and tested on a new design that can either be low or high fidelity. Eventually, the insights can generate a better understanding of how to improve the design requirements (section 4.4.2). 38 5 Planning The chapter outlines a plan to achieve the aim described in section 1.3. Moreover, the chapter presents the work process with a selection of the methods that are planned to be executed and an explanation of why specific methods were excluded. 5.1 Project Initiation Week 1-2 will involve a project initiation consisting of a pre-study and planning phase. The weeks will include introductions to the authority’s organization, meeting with people in the department, and receiving relevant resources for the project. 5.2 Design Thinking Process Week 3-14 will cover the design thinking process consisting of the following phases: empathize, define, ideate, prototype, and test. Each phase has a purpose and a goal that is planned to be reached with different activities and design methods. 5.2.1 Empathize The empathize phase is planned to involve seven methods; literature review, sim- ulation exercise, think aloud, eye tracking, interviews, indirect observation, and questionnaires as illustrated in figure 5.1. Simulation Exercises Indirect Observation Eye Tracking InterviewThink Aloud Cognitive Walkthrough QuestionnaireLiterature Reviews Empathize Figure 5.1: The seven methods with a yellow background that are planned to be used in the empathize stage. Authors’ own figure. 39 5. Planning The phase will begin with a literature review to establish the theory and find suit- able methods for the project. The eye tracking will serve the purpose of gathering behavioral data about how users currently navigate to the e-services, specifically to understand where they direct their visual attention (section 4.4.1). The selected eye tracking software, RealEye (n.d.), uses the in-built webcam in the computer, which makes it possible to conduct studies remotely, and the plan is to recruit 5-10 users. Moreover, the software enables participants to interact with the eye tracking without the presence of the researchers. The initial thought was to combine eye tracking with think aloud. However, a pilot test revealed that it would result in cognitive overload to perform think aloud while simultaneously interacting with the eye tracking system. To avoid this, the plan changed to introduce a questionnaire immediately after the users have interacted with the eye tracking system and receive more attitudinal data. Nevertheless, think aloud will be executed on its own when doing remote user re- search on visually impaired people since it will not be possible to use eye tracking. The idea is to use a video conference software that enables screen sharing and ex- ecute think aloud while the participant uses a screen reader. This setup makes it possible for the researchers to see how the participant interacts with the website. Furthermore, interviews with people from different departments at the authority will be executed to gain knowledge about the existing system and the previous work, findings, and insights. It will also help understand how the different departments are interlinked and how they might be affected by the result of the project. To avoid ’reinventing the wheel’, indirect observation will be used to receive already collected data (in terms of user feedback and statistics) from previous research conducted by the different departments at the authority. To understand the accessibility issues, simulation with Funkify (n.d.) (Chrome- plugin) will be used to simulate how navigation is experienced by people with cog- nition, dyslexia, motor and vision disabilities. Using automated usability testing software instead of manual testing such as a cognitive walkthrough is because the plugin will provide a quick overview of the main accessibility issues. The idea was to perform a cognitive walkthrough with the help of the WCAG guidelines. How- ever, after further reflection, it is believed that a cognitive walkthrough with WCAG would result in findings that have already been found in previous research from the authority. 5.2.2 Define For the define stage, the plan is to conduct four different methods; affinity diagram- ming, personas, scenarios, and design requirements as illustrated in figure 5.2. 40 5. Planning Persona Behavior Archetypes Affinity Diagramming Design Requirements Scenario Define Figure 5.2: The four methods with a yellow background that are planned to be used in the define stage. Authors’ own figure. The affinity diagramming will be performed first since it helps sort and organize the data gathered in the empathize stage. The idea is to write down everything that might be of interest (subjects, ideas, problems) on sticky notes. This might be done physically or digitally depending on how the regulations of Covid-19 will be during that stage. The sticky notes will then be sorted into clusters and categories to find struggles, opportunities, goals, and opinions that can be used for further analysis. Personas will be created as the second step in the define stage, using insights from the previous step. As mentioned in the methodology, personas are helpful in con- certizing goals and needs, making it a suitable way to use the data categorized in affinity diagramming. The aim is to represent different kinds of users from the target group, preferably with different abilities encouraging an inclusive design suitable for everyone. The personas will be created using a mix of demographics from the users involved in the study, focusing on their goals, skills, and behaviors. Scenarios are selected as the third step since a practical way to get value from the personas is to put them in a context. The plan is to create goal-directed scenarios where the personas’ interaction with the future product is described, creating an initial idea of how the product might be designed to match the persona’s goals. Design requirements are planned to be the last step of the define stage. Hopefully, the previous steps have guided the process in the right direction. However, it might be hard to grasp how to use all data in a helpful way. The design requirements are meant to concertize it and create an overview of what the product should do. The idea is to create a list with all the functional and nonfunctional requirements, creating a natural starting point for the ideation stage. 5.2.3 Ideate The idea for the ideate stage is to perform two different methods; sitemapping and sketching as illustrated in figure 5.3. 41 5. Planning Sitemap BrainstormingFlowchart Ideate Sketching Figure 5.3: The two methods with a yellow background that are planned to be used in the ideate stage. Authors’ own figure. Using the requirements from the previous step, the purpose of the site mapping is to explore different structures and ideas of how to order the content. The sitemaps will be of the visualizing type, created digitally in Figma (Figma, n.d.). The goal is to get a joint view of how the navigational guide’s information architecture should be. Secondly, sketches will be made showing how the different views of the product could look. First, many sketches will be created to explore many possibilities. The sketches will be made using paper and pen. Later in the process, the two researchers will discuss the sketches and write down comments, positive and negative aspects of each (annotations), and decide which sketches to use for the prototyping stage. The brainstorming method is not selected as a different method since brainstorming is preferably done in larger groups in its raw form. Only two researchers are carrying out this project, making it hard to implement a ’real’ brainstorming session without external people involved. However, both site mapping and sketching could be seen as a light version of brainstorming since it is a different idea generation. 5.2.4 Prototype For the prototyping stage, digital prototyping (i.e. high-fidelity prototyping) will be used instead of paper prototyping (i.e. low-fidelity prototyping) as illustrated in figure 5.4. 42 5. Planning Digital Prototyping Paper Prototyping Prototype Figure 5.4: The one method with a yellow background that is planned to be used in the prototype stage. Authors’ own figure. The reason for selecting a paper prototype often refers to the implication that it is faster and cheaper to create. However, the authority already has a high-fidelity ready-to-use design system in the program Axure (Axure Software Solutions, n.d.), implicating that it should not take more time to create a digital prototype for this project. Furthermore, due to limited time resources, there is no possibility to create both types of prototypes, and the digital one opens up the opportunity to offer a more finished concept to the stakeholders. Thus, this project will be implemented with a digital prototype created in Axure. 5.2.5 Test The evaluation will cover eye tracking and A/B testing as illustrated in figure 5.5. A/B Testing Think Aloud Eye TrackingCognitive Walkthrough Test A B Interview Figure 5.5: The two methods with a yellow background that are planned to be used in the test phase. Authors’ own figure. The test phase evaluates the created prototype and highlights potential issues to consider in future editions. The plan is to use eye tracking to understand how users behave during the interaction. A/B testing will serve to test the navigation of the final prototype against the current navigation at the authority’s website. In turn, it will also help decide which option is most effective. Furthermore, the results from the evaluation will contribute to the revision of the design requirements. The list will be updated and become a collected summary of 43 5. Planning design guidelines for designing navigation to e-services. In other words, the design guidelines will be based on the research from the thesis project. The methods that will be excluded are cognitive walkthrough, think aloud, and interviews. The expectations are that the findings gathered from both methods will be similar to the ones in eye tracking and A/B testing. 5.3 Report Writing The report writing should be performed continuously during the whole project. Most focus on the writing will be at the beginning (methods & theory) and the end (execution, result & discussion). Notes should be taken after each method is completed, making sure that no details are forgotten. Week 15-18 will cover intense report writing. The weeks 18-20 will prepare and execute the thesis presentation and revise the report to deliver it to both Chalmers University of Technology and the authority. 44 6 Execution and Process The chapter is structured in sections based on the different phases of the design process. Each conducted method is presented under the correlated phase with a description of the execution, a summary of the results, and a box with the most relevant insights to bring forward in the process. Note that all information has been collected in Swedish and translated into English for this report. 6.1 Empathize The empathize phase involved literature reviews, a simulation exercise, indirect observations of existing materials provided by the authority, two expert interviews from the authority, a user study consisting of eye tracking and questionnaire, and a think aloud. 6.1.1 Literature Review Literature reviews were conducted to understand the design topic and theory. The researchers found literature from Chalmers Library, ACM Library and Google Scholar. In addition to that, other credible sources such as websites (e.g. nngroup.com, interaction-design.org etc.) and theses have been accessed. The main keywords that contributed to the findings were: design methods, web navigation, decision making, usability and accessibility. The findings from the literature review can be found in the theory (chapter 3) and methodology (chapter 4). 6.1.2 Simulation Exercise The Google Chrome plugin Funkify was used to understand how the navigation to various e-services on the authority’s website is experienced by people with dyslexia, cognitive, visual, and motor disabilities. The goal was to gain first-hand experience and understand what to consider when making a navigation accessible. Funkify has ten simulators acting as overlapping filters on the screen, allowing the users to experience the same information as they would do if they had the simulated disability (figure 6.1). This section describes the different simulation disabilities, while the conclusions from each simulator can be found in appendix A. 45 6. Execution and Process Figure 6.1: The ten different simulators on Funkify. Authors’ own version of screenshots taken from Funkify (n.d.). Results from the Simulation Exercise The dyslexia simulator made all the letters scramble, creating an overwhelming feeling while reading the text. An example is shown in the figure 6.2. Dyslexia Figure 6.2: A screenshot of the authority’s website with the dyslexia filter. Au- thors’ own version. The vision simulator simulated multiple vision impairments such as color blindness (black/white, red/green and blue/yellow), blurry vision, tunnel vision, peripheral vision, sunlight casting on the screen and lower vision often experienced by older people. The sunlight casting simulator used intensity to highlig