Empowering Independence Through Design: Investigating Standard Digital Design Patterns For Easy-to-Read Users.

As designers and researchers, it is our duty to ensure information accessibility for all, irrespective of cognitive abilities. Currently, Easy-to-Read (ETR) is commonly used to simplify text for individuals with cognitive impairments. Although design aspects of text comprehensibility have recently gained attention, digital design patterns remain relatively unexplored. Our understanding of how ETR users interact with digital media, and how to design specifically for their needs, is still limited. Our study involved observing 20 German ETR users engaging with a digital PDF and a website designed in a participatory process. We collected data on their access to digital media, personal use and workarounds, and their interaction with digital design patterns. Tasks on the smartphone were completed mostly successfully, while only 50% could navigate a digital PDF. In both cases, visual cues played a significant role. Our findings contribute recommendations for beneficial digital design patterns and future research.


INTRODUCTION
According to the 2006 United Nations Convention, accessibility is a human right [52], but the range of users' needs and abilities is vast.Easy-to-read (ETR) is therefore a key to an inclusive society that allows every person to access and understand information [1] and was created to empower participation of people with low reading skills due to so-called intellectual disabilities (ID).The principles of ETR were developed using print media and then applied to digital media, although they are entirely diferent and have unique requirements.In addition, web accessibility research has not yet adequately represented users with cognitive impairments [87], and this is even more true for ETR users.
Although some claim that ETR might support a wide audience like the elderly and foreign language speakers [54], our study focuses on people with ID for two reasons.First, this was the original target audience of ETR, and this group is already heterogeneous regarding age, reading level, experience with digital media, and cognitive and motor skills.Second, there is a scarcity of user research and testing since it is difcult to access this group.People with ID are often under legal guardianship and live or work in institutions for people with disabilities.We concentrate on users with ID who can at least read words.Both our co-researchers and all test participants have a diagnosed ID and individual additional impairments.We use the term 'ETR users' given the community's rejection of the term 'ID' due to its stigmatizing connotations and the inadequacy of their self-chosen term 'learning disabilities' in cognitive science.By using the term 'ETR users', we refer to users who read ETR regularly, regardless of their reasons.
Originally, ETR focused on simplifying language and participatory testing [9,92].Social workers translated content using programs like Microsoft Word, without considering professional visual design [62].Print products and the frst guidelines [8,53] refected this lay design practice.Instead of aiding comprehension, the typical designs were dysfunctional, negatively afected readability, and contradicted design research results [66].Design researchers, especially those with an emphasis on typography, are experts in readability and visual design.Their fndings are currently being incorporated into the upcoming German ETR guidelines, which will then be ratifed as a standard that includes a chapter on visual design [11,22].
In digital media, ETR is increasingly ofered as an accessible alternative.Digital design includes text, design, and interaction elements.These elements are represented in digital design patterns, which are the focus of our study.The term 'design pattern' originates in architecture and describes a solution to a recurring design problem [2].Design research picked up this prototypical (rather than taxonomic) approach, as it corresponds to the way (graphic) design is practiced.For example, a headline on a page stands out as the primary piece of information because the type is usually larger and bolder than the running paragraph text of the articles.The reader intuitively applies a reading strategy of skimming headlines on a page and then choosing an article to read.Thus, design patterns work because they directly correlate to our perceptual system, not simply because they are learned [86]; it is nature, not just nurture.In digital media, code and functionality are embedded in the patterns, e.g.selecting the 'hamburger icon' (usually represented by two or three horizontal lines) opens a menu.Most digital design patterns are established, familiar, understood intuitively, or make use of accepted mental models.New patterns, however, need to be learned and adopted.For example, the swipe gesture was not familiar prior to its usage on mobile devices [45].While technology standards have been explored, the role of design in ensuring the universal usability of digital design patterns still awaits research and testing [68].
Given the scarcity of published research fndings on digital ETR use, it is unclear how ETR users interact with existing 'standard' digital design patterns, despite the vital role that such patterns play in making digital information accessible and comprehensible.Therefore, our paper investigates the following research questions (RQ): • RQ 1: How do ETR users operate digital devices and what limits the user?• RQ 2: Does the target audience understand and employ digital design patterns?• RQ 3: Which digital design patterns might be helpful for ETR users?
In this article, we summarize the views of diferent disciplines on the topic, outline our interdisciplinary and participatory project structure, present the results, and, through discussion, propose an optimized user interface design for ETR users.

RELATED WORK
We will begin by reviewing social science research on ETR users' media use, followed by an exploration of User Experience (UX), accessibility guidelines, universal design, and research in cognitive psychology, linguistics, and design related to ETR.Finally, we will analyze the trend of individualized interfaces.

Usage of digital technology by ETR users
Access and the ability to interact with digital media, such as web pages or accessible PDFs, are key requirements for modern societal inclusion [26].Two meta-studies-one conducted in the EU in 2017 and another in the UK in 2016-revealed that only 50% of people with ID have access to the internet, compared to 70% of the general population [60].When they engage in social media they have positive experiences in terms of friendships, self-esteem, and enjoyment [18].Identifed barriers encompass concerns about safeguarding, literacy, communication difculties, and accessibility issues including inadequate equipment [18].A 2020 Norwegian study found that touchscreen technology can provide empowering opportunities for individuals with ID, even if they encounter challenges in reading, writing, or using voice-activated strategies [67].Furthermore, a 2018 Australian study highlighted how young adults with ID engage in social media based on personal interests.Nevertheless, their participation is hindered by issues such as low literacy, authentication problems, advertising content, and a lack of awareness regarding tools that could address these issues, such as voice search, auto-login, password retrieval protocols, and ad blockers [6].A 2021 study in Germany suggests an internet access rate of 93% [7].However, as this survey was conducted online and required internet access to participate, it is questionable how representative these fndings are.Access to the internet might be discouraged by caretakers or guardians with little digital skills, who see more dangers than opportunities in internet use, and assume that people with ID cannot, with appropriate concern and choices, classify viruses, Trojans, spyware, and chatbots, or may incur debt when shopping online [12].Regarding gaining digital skills, there is a lack of training for staf and residents in institutions [13,51,91].Practice showed that ETR users are in principle able to acquire digital skills [49] and gained far more competence and knowledge than was expected.Unfortunately, existing digital competency tests (such as [19] or [3]) are not suitable for ETR users since the questions are too difcult from a language and cognitive point of view.ETR users' digital competency is therefore difcult to measure and largely unknown.

UX principles and recommendations for accessible design
Besides users' digital competency, digital products must be designed according to user needs.General UX principles defne quality components for usability, best practices in user experience design, and web accessibility [36,38,40].The Web Content Accessibility Guidelines (WCAG) guidelines [89] follow general UX principles.Seemann and Lewis [2019] summarize two studies [32,69] and conclude that "the problems encountered by many people with cognitive disabilities are, broadly, the same usability problems that afect all users, but the impact on people with cognitive disabilities is more severe" [61].The WCAG are further specifed by publications of the Cognitive and Learning Disabilities Accessibility (COGA) Task Force [82].The authors state that digital inclusion for the COGA group is far behind other groups of disabilities and there is a lack of user research and user testing [61].Practice-oriented contributions like the 'a11y project' [57] provide valuable resources in which UX designers with neurodivergent backgrounds contribute based on their life experience.Meredith identifed "access" and "success" as keys, suggesting that "abundant design" [50], which provides the user with two (or more) ways to achieve the same goal, would facilitate a better, more successful user experience.ETR users are a subgroup of the COGA group and difer from other groups of users in that dyslexia or ADHD is not the predominant reason for low reading skills.

Universal Design and touchscreen devices for users with ID
Universal Design (UD) looks at designing for the greatest number of possible users: it promotes the idea of designing products and environments to be used by as many people as possible, accounting for as wide a range of abilities as possible, in the greatest number of scenarios as possible.[81].For users with ID, the main benefts are being able to use the same product as everyone else, which avoids social stigma and extra costs, gives them better opportunities to interact with (non-disabled) family and friends, and to access the same digital content as everyone else [67].An example of UD on smartphones is the accessibility features: high contrast backgrounds and text magnifcation have been built in to assist users with visual impairments, but are also used by many people without visual impairment.Studies have shown that touchscreen devices are easier for users with ID [59,67] since errors can be overcome and the user experience is more comfortable than using a keyboard.One study from 2016 and another from 2018 reported difculties resizing and using two-fnger gestures for zooming.The tests were conducted using 'artifcial' UI designs created to test tasks on an iPad Mini [59] and a smartphone [88].In the smartphone study, the instructions looked like a list or menu, leading participants to tap and expect another page.These fndings were interpreted by Williams and Shekar [2019] as "priority of appearance" and they concluded that "the more visceral the afordance, the higher the usability appears to be." [88] However, we attribute these difculties to misapplied digital design patterns.

Cognitive psychology and research on ETR in print
UX principles refer to fndings from behavioral and cognitive psychology [90].Chandler and Swellers' [1991] cognitive load theory [20] distinguishes the Intrinsic Cognitive Load (the efort associated with a specifc topic), the Extraneous Cognitive Load (the way information is presented), and the Germane Cognitive Load (the efort required by the learner to understand the material).ETR users might have less previous knowledge and, due to their lower cognitive skills, face a higher Germane Load.Consequently, they would need materials with an adjusted Extraneous Load, hence an 'easier' language and design.This is in line with psychological research into ETR suggesting to consider the recipient and the text side [21].Key recipient issues identifed by Christmann [2017] are low retention [44], difculties with inference, e.g., linking pieces of information in a text and with previously learned information [15,16], and problems in suppressing irrelevant meanings [29].On the text side, ETR users are helped by linguistic simplicity, semantic brevity and continuity, cognitive organization, and motivational stimuli [21].However, these linguistic aspects do not take into account the visual design of the product and its impact on text comprehension.In a joint linguistic and design study on ETR, Bock & Sieghart [2019] added the design factors 'legibility', 'readability', 'structure', and 'motivational design' that facilitate text comprehension for readers using ETR [65].

Reading research in general and into ETR
Design research provides information on how to visually support reading.Typography (the design of a text) plays a role in several phases of reading: perception, assigning meaning, and comprehension.The frst phase is the perception and decoding of letters, which is the subject of legibility studies and is addressed in nano and micro typography [4].Letterforms designed to match the physiology of the retina and visual cortex facilitate reception [75].A 2022 study involving users without disabilities examined fonts in digital media and confrmed that "diferent fonts are efective for diferent people", noting that some fonts could increase reading speed by 35% [83].This might also apply to ETR readers who, in a study with print material, read the same typefaces faster than non-ETR readers-compared with their generally lower number of words read per minute [64].The second phase is to assign meaning to what is seen.Prototypical design using established design patterns gives an indication of what kind of text to expect, even without reading a single word [78]; a novel looks diferent from a poem, a newspaper, or a contract.Macro typography creates recognizable structures and layouts using linguistic text types and genre-typical design.Contrary to previous assumptions that ETR users do not know genres "because they do not use them in daily life" [14], an interdisciplinary study found that ETR users have knowledge of linguistic text types and design genres and use the visual language to understand the content [10,63].The third phase is content comprehension.This is dependent on the reader's prior knowledge, cognitive abilities, reading medium, and environment [41,58,75,77,85,86].Typographic systems are highly complex, with many interdependent factors.Expertise is required to select optimal fonts and specify micro-and macro-typographic settings in order to create appropriate design patterns and set up valid studies on this topic [5].

Individualized interface design and tools for ETR users
The benefts of personalization (done by the system) and customization (done by the user) have been discussed for some time [55,56] and eforts have been made to automatically confgure personalized accessibility features.An example is the 'morphic' tool, which has been developed with personas and tested with students who have an impairment, but also a much higher level of (digital) literacy than ETR users.While the technical parts have been solved, Vanderheiden [2020] claims that "we have not yet found the approaches needed to reach many of those who could use these features" [80].
The WCAG experts see opportunities in adapting websites to the individual needs of the COGA group and, as a bridge, suggest tools to improve access [61].Currently, several digital products are developed to adapt content and design for ETR users.Some tools will not change any visual design or interaction design and focus solely on text simplifcation, such as 'capito digital' [17].'Easyreading' [24], a product to improve the cognitive accessibility of digital documents "by personalization through annotation, adaptation and translation", [25] currently lacks user tests [34], which are planned in the follow-up project 'EVE4all' [24].From a UX design point of view, the aforementioned WCAG and UX standards might not be met and the operation might involve efort on the user side (and thus customization).Another tool uses state-of-the-art UX design, but fails to consider the low reading skills of the target group: The plugin 'accessiBe' [23] works with extensive settings using standard language.The average ETR reader would have to cope with an unmanageable amount of text.Furthermore, the settings neglect the hidden dependencies of typography and perception laws: font choice, type size, and word and line spacing are mutually interdependent, as well as contrast, color of type, and background.The tool allows dysfunctional combinations to the point that completely unreadable interfaces are displayed.There are several aspects to consider for individualized interfaces.First, users are not typography experts.In a study involving users without disabilities, Wallace [2022] found that the typeface which was fastest read did not match participants' preferences, which might mean that "people do not know what is good for them in terms of font choice for reading" [83].This is refected in a study of ETR users: Preferences did not match reading speed [64].Second, UX research has found that most users stick with the default settings rather than changing them [39].Third, reading is not the only aspect to be considered, as Nielsen [2022] remarks, "Performing tasks on websites involves reading, but it also requires many more advanced skills, such as navigation, searching, and the ability to judge and make decisions based on what's being read" [37].Fourth is the technical aspect: Many custom applications do not support standard, system-wide accessibility functions such as zoom, contrast modes, and keyboard controls [43].
In summary, the specifc group of ETR users is poorly researched.We do not know what kind of devices they use or how much access to digital media they have, let alone whether the products currently being developed will work for them.ETR design research in print does not take into account digital media and interaction design.Some digital ETR products lack expertise in user interface design and typography, while others lack knowledge of ETR users' needs.This paper contributes to closing these research gaps by focusing on digital design patterns and empirically testing ETR users' interaction with them.

RESEARCH METHOD
This study was part of the LeiSAparti-study [79], an interdisciplinary participatory research project at the University of Cologne (Germany).The team consisted of three linguists, three social scientists, one designer, and six co-researchers (see fgure 1).The co-researchers were recruited from a university program [76] dedicated to engaging individuals with ID in research projects, utilizing their real-life expertise.All six co-researchers had a cognitive disability and a wide range of additional individual impairments.While most were capable of verbal expression, one participant required translation assistance from a social worker.The multidisciplinary nature and previous ETR experience of all team members strengthened the participatory process, the writing, and design of the test materials.This resulted in fve sub-studies, of which the design study is presented here.This design study is the frst study with digital media in a series of studies aimed at identifying how design can contribute to greater inclusion of ETR users.Its main contribution is to explore ETR users' experiences with digital devices and their use of digital design patterns.It builds on and complements previous evidence from printed media.Therefore, we purposefully created realistic test materials to raise the ecological validity of the study.The creation of these materials was entirely guided by the participatory design process mentioned above, with the notable inclusion of co-researchers using ETR.Although this approach allowed us to acquire novel insights into ETR users' interactions with digital media, it also limited our possibility to systematically compare diferent digital design patterns for identical purposes (e.g., whether gesture-based zooming or a dedicated zoom menu is preferable).We will return to these refections in the limitations section.

Test materials
The test materials were designed through a participatory design process (PD).The co-researchers opted for a website, while scientists also wanted PDFs that could be forwarded or even printed.The coresearchers favored a 'scientifc' looking design and selected a clean user interface design based on several examples.In particular, they disfavored an ETR design solution with childish illustrations.In an iterative process, we reduced the information density of the PDF and HTML pages and took into account the specifc needs of the co-researchers, such as limited attention span and tendency to lose track (structured and clear layout), low vision (high color contrast and enlargeable type), color sensitivity (use of only one highlight color per chapter), and operation of the smartphone with the nose or with motor impairments (large enough buttons and more space separating clickable areas).The fnal design was coded according to WCAG accessibility guidelines for testing and is available as supplementary material.
Operating a digital interface successfully requires users to perform tasks involving digital design patterns.A standard website or an interactive PDF contains numerous embedded patterns that facilitate various navigation and usability strategies: • Scrolling (Observe whether the interface design of digital media is understood, where the user scrolls instead of turning the page) • Zooming (Examine the use of the magnifying glass icon in the navigation bar in the PDF, or the two-fnger pinch gesture on the smartphone) These digital design patterns were selected for our test as they turned out to be useful for tested materials during the participatory design.The frst questionnaire confrmed that the test participants did use standard websites and thus might be familiar with standard digital design patterns.Accordingly, we formulated tasks (see table 1) along an expected test scenario, requiring the participants to either use the defned digital design patterns or fnd alternative approaches.The rationale was to assess the utilization of patterns and their impact on successful interaction.We designed two variants to allow testing an visible menu (variant A, see fgure 2 against a navigation overlay (variant B), see fgure 3.Both variants are available as supplementary material and can be accessed via these links: variant A [47] and variant B [48].

Procedures
The test consisted of three on-site sessions.In the frst session, ethical consent was obtained and the reading test and sociodemographic questionnaire were completed (including data on access, digital devices, applications used, personal settings, perceived difculties with digital media, and support).The questionnaire on   PC use and visual knowledge of PDF navigation bar icons was ad-as supplementary materials.The choice of test devices was based ministered in the second session, prior to testing the PDF on a PC. on the participatory design, the results of the frst questionnaire, In the third session, we administered the third questionnaire asking and data from the Swiss Ministry of Health [74].We attached a about the visual familiarity of the smartphone icons in a schematic PC keyboard and a Logitech optical mouse to our computer and drawing and thus in context, and performed the test on a smartselected a Samsung Galaxy A12 for the smartphone tests.These are phone.The questionnaires were designed in ETR and are available

Ethical approval and ethical consent
Ethical approval was granted by the medical faculty's ethics committee of the University of Cologne, reference 20-1591-1, 26 July 2021.For our participatory approach, it was crucial that people could choose to take part in the study, have rights over their data, and be able to withdraw, regardless of their legal status.Hence, we developed a consent form in both ETR and standard language that meets legal requirements and is accessible to ETR users and their legal guardians.

Data measurement
Based on a previous study, the 'lea diagnostic tool' was chosen to determine reading level [30,31].Quantitative user data was collected using the three questionnaires described above, and testing was conducted using read-aloud tasks as described in table 1.Additional qualitative data was gathered through participatory observation and structured interviews.While questionnaires only capture reported behavior, participatory observation documents behavior in vivo, and interviews capture data from the participants directly after the interaction [28].We decided to keep the test situation as unintrusive as possible, and thus only recorded audio of the tests since, in previous studies, ETR participants expressed discomfort with videotaping.This also reinforces the validity of the study.

Test Participants
We collaborated with a facility for people with disabilities [70] to recruit 20 ETR users (16 men and four women, between ages 19 and 60, average age 33.4).The majority (17) lived in the institution, two with their parents, and one in assisted living.None of them had any school certifcate.Sixteen participants work in workshops for people with disabilities (metal, gardening, agriculture, home economics).The remaining four participants are unable to work and are in the support center.All participants have diagnosed ID along with individual additional impairments.Caretakers assisted in selecting individuals with low reading skills who could envision using a digital device.Our subsequent reading test specifed their reading abilities, which often difered from the caretakers' estimations, as evidenced by 'lea' results.'Lea' reading levels range from 1 to 6, with 'lea'-level 6 corresponding to level 1 of the International Adult Literacy Survey [30].The majority of the participants had 'lea'-level 3 or 4 (see table 2).Level 3 readers could read individual words in the context of sentences and follow simple instructions, especially with images.Level 4 readers could pick out individual words from a text, recognize structures of simple forms, and extract one to two directly contained (verbatim) or indirectly contained pieces of information from short and simple texts with explanatory images [30].

Data analysis
We used the grounded theory (GT) method by Strauss and Corbin [72] as it allows combining qualitative and quantitative data [35] and takes advantage of the theoretical sensitivity a design researcher contributes by professional experience [72].In the quantitative analysis, all questionnaires, task results, and audio recordings were analyzed and phenomena were extracted.The phenomena were further enriched using the qualitative GT method, which codes full transcripts in three stages, alternating between collecting and analyzing data until theoretical saturation is reached.Strauss and Corbin [1996] state "saturation means that no additional data are being found whereby the sociologist can develop properties of the category" [72].In the frst stage, codes describing user behavior were extracted and categories were synthesized (open coding).Secondly, multiple coding paradigms were used to organize the categories (axial coding).Finally (selective coding), core categories were identifed that explain the phenomenon under study [71].After analyzing 12 transcripts from six participants, we found that we had reached theoretical saturation.We were able to describe in depth the core categories 'access', 'user side', and 'product aspects'.The user side consists of personal variables such as reading level, digital literacy, visual knowledge, and user strategies.The product side is characterized by visual design and technical accessibility.

RQ 1: Usage of digital devices and limitations
The institution only provides Wi-Fi access at the caregiver's PC within the residential groups.Nevertheless, 17 out of 20 people (85%) stated they have access to the internet and 85% own a smartphone (see table 2 for details).While 50% said they had access to a PC in the questionnaires, the interviews revealed that this is limited to Table 3: Assistance, perceived difculties, change of system settings on own device, online orders, and downloads.
occasions when their family or carers shop online with them.In one case, the machine at work (TNT) was considered a PC.Only seven participants (35%) got help from friends or family (see table 3).Ofcial training is scarce or even not encouraged as one participant without their own smartphone reported: I said I wanted a smartphone.Sarah gave me hers and said, So, now you write.[. . .] I tried something and almost deleted the phone (ME-3:314).Only participants who live in assisted living receive training.This enabled a user at the age of 60 who received his frst smartphone six months prior to acquire skills in a short time.They could not use the PC, but were trained enough to pass a smartphone test: It's good that this assisted living project exists.We are looked after twice a week by people who are trained to do this (10:572).
External expectations, advice, and bans often signifcantly impact the participants' activities and experiences using the internet.This is refected in the emotional responses to the answers about whether participants download fles (55% never) and order online (75% never), as well as in the tests.The risk of falling into debt was conveyed by the tutors-sometimes with drastic examples (like stealing for participant 27C)-and internalized or sometimes bypassed, as the following example shows.The participant vehemently refused to download a PDF on their device and explained: No, no, because I have a partner contract with my mother.[. . .] my mother pays THE contract.When I download something, and it's the wrong thing, then my mother. . .pays more money.That's why I don't download anything.Except for free apps.That's what I download. . .but I read through it frst (MC 14:334f).
When the ETR users commented on the personal use of their smartphones, the applications installed on their own devices were standard applications chosen according to personal interests.For example, one participant described a farming game, where they can overcome the limitations of their work environment and make independent game-related decisions: It's so much fun.This is my fence that I built myself (C31: 459).For others, the smartphone enables privacy (10) or exchanging afection (27C).Despite reading difculties, only two persons had a text-to-speech-reader installed, 40% changed the system settings to a larger font, and no one used other assistive technology.This confrmed our decision to investigate the use of standard digital design patterns.Difculties reported were fnding and understanding information, technical issues, pop-ups, ads, and bugs.Six participants (30%) found nothing or nothing at all difcult.  of a magnifying glass circle-and-stem with a nested plus sign on a PC interface.In the second questionnaire (see fgure 4), we asked about the familiarity and representation of the icons.In contrast to our co-researchers, only 45% of the testers correctly described the zoom icon as an enlarging tool, and the hand icon was often read as a literal hand or even as a stop sign, (or interpreted in line with the familiar TNT machine user interface as manual mode).
The hamburger icon was somehow familiar, but only 10% had a vaguely correct association (mark, load images).The other 40% made a (wrong) literal interpretation such as lines to sign or something from math (equals sign).Even in the third questionnaire, where the icons were shown in context, the hamburger icon was only correctly classifed by 15%.Spatial and contextual memory seemed to play a role in the 30% of wrong answers that associated the hamburger icon with settings, battery status, Wi-Fi, or camera, functions that are usually displayed in the top right-hand corner.However, actual menu use in the test was much more successful than the isolated visual query suggested: 90% clicked on the hamburger icon and operated the menu, but when asked what it was, only 25% could correctly explain.As an entity, the functional design pattern 'menu' appears to be familiar (see fgure 5).

Operation of PDF on PC:
Unexpected problems.The PDF on the PC proved to be much more difcult for our participants than for the ETR co-researchers involved in the participatory design.Regardless of their reading level, 50% of the test persons were unable to complete the tasks, because, for example, the PDF software froze (in four cases).Some participants accidentally opened windows such as additional tools, developer tools, or shortcuts, and could not close them.On smartphones, however, most participants were able to complete the tasks.

4.2.3
Operation of digital design paterns on smartphone.All participants were able to scroll and select links on smartphones.Some had motor difculties with selecting, though those did not impede them from solving the tasks.The close button was familiar: 80% used it and some could describe the icon.Indeed, having a dedicated close button turned out to be vital for interaction design: 80% tapped on underlined words to open explanatory contextual menus.Difculties in closing the window without a dedicated close button showed up: 45% clicked on the background, 30% scrolled, and 15% failed to close the window.The back-to-top button was less familiar-only 45% used it for navigating directly back to the top (see fgure 6).
Regarding the visible menu or navigation overlay (see fgures 2 and 3 and for results fgure 6) there were no clear preferences for any variant.The main arguments in favor of variant B were the ability to reduce the text, while variant A has the advantage that everything is immediately visible and participants expected fewer operational problems.Yet we observed that the prominent button with abundant design was used by 85%, while the icon at top left was only used by 10%.

RQ 3: Indications for helpful digital design patterns
We observed several strategies of the co-researchers and participants that may be able to guide the design of digital design patterns.

4.3.1
Reading on the smartphone is perceived as easier.Large amounts of texts overwhelm ETR users, as this response to screen 7 in fgure 3 shows: «anxiously resistant» Should I read the whole thing?(MC14:239).Learning to read on paper was often a frustrating experience: I hate reading.I don't go to school anymore (M87:288).
That being said, reading on the smartphone was more fuid, and one participant stated: I'm really good with the phone.It's the easiest for me than using the computer or writing or reading on paper (MC14:150).Digital reading can be an opportunity for ETR readers to improve reading, as it is associated with leisure and is less biased.
In addition, written text has further qualities, as the following examples show.First, it creates lasting visual memories.One might think that text could be replaced by audio messages to avoid strenuous reading, for example in WhatsApp communication.However, even at reading level 2, text messages to loved ones were used in conjunction with emoticons and audio messages and were proudly presented (27C).Another aspect is to feel included.One participant stated they did not want to receive any diferent emails than their sports colleagues without impairments do, and rather extracted the important information (names and dates) by skimming the difcult text (C31).

Strategy: Orientation on visual hints and spatial memory.
Co-researchers and test participants relied heavily on visual cues and spatial memory, sometimes to avoid reading.One participant scanned the content list of the PDF (see fgure 1, test material PDF on desktop) for the frst letter (in this case, 'L') and navigated extremely fast to the content (in this case, chapter '4.Lust auf Lesen').They explained their strategy: Lust... you said 'L' and zzzzt «pantomiming how they then found the word on the screen that starts with L» -I saw it and I knew that's it (C31:297).During the interviews, they explained relying on spatial memory for interacting with a menu: I go there, «counting of» one, two, yes, in the third (C31: 470f).They claimed more reading wasn't necessary when strong visual hints exist (such as lists with various word lengths or that are numbered) or when spatial or contextual memory allows them to navigate to menu items rather than unnecessarily reading all text.
Similarly, another participant's favorite application was a clothing brand for construction workers [73].The state-of-the-art UX design caters to the low reading skills of presumably foreign-language speakers and reinforces all entries with icons (e.g.delivery status with an icon of a van).They read otherwise difcult words fuently since they knew the icon and its meaning, such as 'Warenkorb' (shopping basket).

Strategy: Selection of text via image-headline combination.
The headline-image combination appeared several times during the interviews when showing personal favorites.For example, it was used to fnd news about a rapper by using the image search in Google: This is Bonez MC (MC14:463) and then reading the headline and choosing the article.The picture and the short text gave them preliminary information.The design pattern 'headline' was familiar and used for orientation within texts: Where do I start reading?With the headline?(MC14:212).After confrming, they read by starting with the headline (see screen 7 in fgure 3).

Strategy: Clever combination of the functionality of two apps.
We have often observed the smart use of (limited) digital knowledge.For example, one participant (reading level 2) showed a strategy during the interview to motivate their father to download a flm.In a frst step, the participant typed the movie title into Google with a spelling mistake ('Soma' instead of 'Sommer').The application showed text-image combinations and, in a second step, entries with corrected spelling.Consequently, they could see the 'Soma Festival' and the 'Sommer Film' with images, and chose the image they recognized.They forwarded the link with its attendant text they themselves would never be able to type through WhatsApp.The interaction went smoothly: On Google, so there was nothing difcult to fnd (27C-3:181).With their limited digital knowledge of two applications and some persistence to overcome false results, they achieved their goal.4.3.5Text reduction strategies.We observed various strategies to reduce the amount of text to read.40% used the smartphone settings to produce larger type size on their device.In the smartphone test, the two-fnger pinch gesture for zoom was often used when reading became tedious.At the same time we observed that small type could be well read (e.g. in folder names) and therefore assume text enlarging is a strategy to reduce the amount of visible text, and thus the perceived extraneous cognitive load.This is in line with an observation on plain language, where users preferred large type, not because they could not read small type, but because it made them feel there was less text on the page to cope with [84].Another strategy was to revisit texts, breaking the reading process up into several steps: I am usually too lazy to read.(M87-3:235).Sometimes I give up.[. . .] When I feel good again, I keep trying (M87:299f).

Smartphone allows self-determined participation, communication, and media consumption
Our results show that ETR users practice self-determined consumption of movies, music, recipes, sports, and celebrity news on their personal smartphones.To quite an extent they are able to become part of their respective digital communities, participate in ageappropriate activities (dating, gaming, gossip), gather information (news, images, travel and mobility), and communicate with family, peers, and carers.Our participants value and beneft from the use of applications and are highly motivated to overcome challenges, which is in line with the fndings of previous studies [33,67].ETR users are already active in the online world and our data on internet access show that the strategy of keeping those users away from the internet by means of Wi-Fi restrictions is not feasible in the age of smartphones.Sadly, our research reafrms that the long-requested digital skills training for individuals with ID, which has been advocated for decades [6,33,51,91], is still not being implemented.

Digital skills are built using standard applications on personal devices
For ETR users in institutions, most of the learning is self-guided on personal devices, usually smartphones.During their daily use of smartphones, they acquired knowledge of standard digital design patterns.As a result, interaction on smartphones (such as scrolling, zooming, tapping, and menu operations) was mostly successful.In contrast, zooming the PDF on the PC (clicking with the mouse on the zoom icon in the navigation bar) was not a completed action for most.This is consistent with previous research with individuals with ID showing that touch devices ofer greater usability compared to a keyboard and mouse.[59].There were no difculties with the two-fnger pinch gesture for zooming or resizing, in contrast to previous studies [59,88].When a pattern was missing (such as the close button in the word explanation window), ETR users would either switch to alternatives or be unable to continue.Less common patterns, such as the back-totop button, were not familiar.The new icon for overlay navigation was understood with the additional textual information, but only 30% reused it during the test.The hamburger menu was used, even though the hamburger icon could not be interpreted correctly.These results may suggest three things: ETR users have internalized this pattern (even if they cannot express the meaning); that the design of the digital design pattern 'hamburger' follows a mental model; that it adheres to the perceptual system; or any combination of these three factors.

Design issues on the product side: Interfaces difer in extraneous and intrinsic cognitive load
The more successful test on the smartphone may also be related to the design and the cognitive load it imposes.When we analyzed the designs of our test materials within the test devices, we realized that several factors increase the extraneous cognitive load of PDFs.
There are many control bars on a PC and PDF interface which unimpaired users consciously ignore.The ETR participants could not do so, and multiple control bars with additional functionalities were displayed (by default or by accident).On a smartphone screen, there is much less text due to the smaller screen size, and, on newer software and in our tests, the control bars disappear as soon as the user scrolls, reducing the extraneous and intrinsic cognitive load: there is less information to read and digest, and fewer tasks to choose between (see fgure 7).Another observation was that the accessible PDF navigation bar was not used by any of the participants.This contradicts the general assumption that accessible PDFs are barrier-free.This was not the case for our ETR users, for whom the additional control bar likely increased the cognitive load.

Recommended digital design patterns
Back in 2003, Johnson and Hegarty, in their study with young people with ID, discovered that website accessibility was not the main problem (or could be solved by training), but that websites should be more "graphics based".They concluded [2003] that "what is needed is good website design-the creation of interesting content, with good multimedia content and efcient programming" [42].Since then, accessible programming has improved greatly and ETR users train themselves using their personal smartphones.This strongly suggests that we should be thinking about smartphone optimization and attractive, engaging, and motivating design (as opposed to many of the tools currently being developed).However, insufcient in-depth research has been done on the visual design aspects that help ETR readers.Our study provides compelling evidence regarding the digital design patterns that ETR users learn and fnd useful, suggesting opportunities for further improvements and research.We propose a two-step process in line with the fndings of cognitive psychology.First, the intrinsic cognitive load should be decreased by reducing the tasks and steps required for interaction.Our study demonstrated how this is possible by investigating ETR users' needs in a participatory design process.Second, the user interface design should be created with a low extraneous cognitive load and the digital design patterns should be optimized accordingly.Both steps are mirrored in the key learnings of a 2022 study that suggested "reducing the number of felds to complete" and "includ(ing) more images and icons to make it more appealing" [27].For the second step, we suggest taking advantage of ETR users' sensitivity to visual cues.Based on the interaction strategies of ETR users observed in our study, the following general principles should be followed: • Reduce not only the complexity but also the amount of text.
• Create a clear structure and hierarchy by using the intricate possibilities of typography (such as headlines or image-headline combinations and icons).• Consider spatial arrangement (within lists or across interfaces).• Enrich lists with icons to explain content and provide a second path of access (abundant design).
Some design choices, such as the cart icon added to the word 'shopping basket', may ostensibly increase the extraneous cognitive load, but the ability to fnd the correct entry in two ways gives the user confdence rather than relying solely on their poor reading skills.Successful digital design patterns can have a complex typographic design, yet produce cognitive simplicity.To further investigate how design can improve ETR users' interaction with digital media (beyond the general principles above), we suggest that design researchers improve and test standard patterns or create new ones.The visual design is not the only factor that can be improved.The desired focus on smartphone interfaces opens up new possibilities, such as incorporating auditory or sensory cues like vibrations and sounds to confrm actions.It also has limitations (e.g.no mouse-over).Our study has highlighted the potential of the following, specifc digital design patterns.Consequently, there is a pressing need for further research to delve deeper into optimizing these patterns for ETR users: Therefore it could be advantageous to explore the feasibility of 'show and hide' functionalities to mitigate increased cognitive load.Additionally, investigating features like textto-speech capabilities might prove benefcial.

Limitations of PDFs, participatory design, and customization
Training on a device is key to developing motor skills such as scrolling and clicking a mouse.While the ETR co-researchers were able to operate a PDF on a PC, 50% of the participants failed.Other barriers to the use of PDFs were the strict ban on downloading and the inherent complexity of the design.Although PDFs are the current standard for accessible communication and have the advantage of being accessible for the blind and printable for people without internet access, they are difcult for ETR users to access digitally.
We learned about the limitations of participatory design.The ETR co-researchers had their own PC in their workspace at the university, received training, and were therefore no longer comparable to other ETR users.This limitation goes beyond our study context because it refects the real situation of ETR production.Most ETR text producers work with the same testers for years, who may build up stronger competencies over time than other ETR users.This further emphasizes the need for extensive testing.
Our fndings opened our eyes to the limitations of customization.We realized that new 'special' design could be problematic for ETR users.They have either learned standard digital design patterns already or make use of task-based intuitive comprehension.In the tests, less familiar patterns (such as the 'back-to-top-button') were often not used.In addition, customization will most likely increase the cognitive load, since users must make additional choices and perform additional tasks.

Limitations of the study
Our realistic test materials include diferent design patterns and allow for a variety of usability scenarios and interactions.Although this setup allowed us to acquire novel insights into ETR users' interaction (i.e., their use of digital design patterns), its limitation is that we do not have controlled settings with designs that systematically variate one factor.However, we considered it necessary to frst understand whether ETR users can manage (or even beneft from) the complex patterns within a 'standard' interface.Based on these insights, research can systematically investigate further variations of digital design patterns.
Another limitation is the comparability of PDFs and HTML pages.We aimed at investigating digital design patterns on both types of digital media in a realistic environment.We therefore opted to test PDFs on a PC and HTML pages on smartphones, based on our ETR co-researchers' typical usage.However, we did not expect the situation of the test participants to be so diferent from that of our co-researchers, who regularly used computers.In the questionnaires, 50% of the respondents said they had access to a PC.The qualitative data showed that 'access' often meant sitting next to an operator but not training on the desktop.Consequently, their digital skills on the PCs were much lower than on the smartphones, which prevents comparing results from PDFs and HTML pages.For the present paper, we concentrated on digital design patterns for smartphones for this reason, but we think it would be worthwhile to systematically compare and fne-tune the same digital design patterns on both PCs and smartphones to gain a deeper understanding of ETR users' needs.
The study was conducted with German-speaking participants.Further research could look at other languages and show the overarching efects of design.
We would also like to investigate non-intrusive 'visual' recording methods.For our next study, we are experimenting with screen recording to record success times and button sizes for motor impaired users.

Practical output of the study
Our fndings led to the fnal iteration phase of the interdisciplinary research project.Linguists and social scientists recognized that text comprehension, which was positively tested, was not the only criterion for using the product.The amount of text exceeded the attention span of the test participants and even if they could understand the language, they were unwilling to read the amounts presented.Therefore, we reduced the length of the chapters to a maximum of three screens on a smartphone (up to 80% reduction).The PDFs retained the text length and were designed to make a book of 12 chapters that could be printed and read in small chunks.The fnal website and PDFs can be accessed via this link [46]

CONCLUSION
In our modern digital era, where rapid information exchange occurs through digital devices, ETR users gain enhanced opportunities for societal inclusion through the use of their smartphones.Regrettably, digital training of ETR users for smartphones still barely exists.Findings from the German ETR community align with this notion, showcasing a preference for touchscreens due to their userfriendly nature, ofering straightforward spatial arrangements (in comparison with desktop and mouse control), and an obviously partly intuitive navigation approach across various applications.However, there remains a signifcant gap in research focused on optimizing digital design patterns specifcally for ETR users.Our design research provides evidence that certain digital design patterns empower ETR users by ofering guided strategies that foster greater independence.
Improving functional independence for ETR users should be addressed from both the user and the product side.Regarding the user side, addressing limited access and digital training is crucial.A noteworthy fnding highlights self-directed smartphone training, demonstrated by successful gestures and often successful usage of standard applications.Necessary improvements on the product side, such as optimizing content and visual design within digital patterns, are highlighted by personal strategies for overcoming device and design barriers.Design's pivotal role in augmenting comprehensibility and accessibility for ETR users is evident, gleaned from their signifcant reliance on visual cues and spatial arrangement.Insights that inform the enhancement of digital design patterns are key fndings of our study.The challenge is to develop test materials that maintain ecological validity and incorporate enhanced designs, integrating visual aids such as icons while keeping cognitive load to a minimum.Achieving this balance is imperative for efectively supporting ETR users.
For optimizing digital design patterns (visually), the involvement of design research is crucial, encouraging designers to use their expertise, design insights, intuition, and creativity to improve the user experience for ETR users.Several technical and visual options exist that require designers' input, as described in the aforementioned recommendations.Developing digital design patterns is a multifaceted challenge, as it involves structured thinking in ofering hierarchical text-image settings and triggering reading strategies.This complexity highlights how important it is for other disciplines to also harness the skills of designers.
An inclusive approach to ETR interface design requires two key aspects: listening to the real needs of ETR users and fostering collaboration among specialists in copywriting, design, and accessible coding.Because needs can vary signifcantly, stakeholder involvement from the beginning is crucial in defning needs and restructuring interaction tasks.In summary, optimized digital design patterns facilitate reading and interaction, and can serve as a basis for personalized ETR tools.They also beneft all users with their intuitive simplicity, limited text, clear hierarchical structure, and multiple task completion paths.
In our next study, we aim to create diferent design iterations using digital design patterns derived from those previously mentioned.We will test specifc scenarios that enable independent choices, such as selecting events, confrming dates, or making particular requests.

Figure 1 :
Figure 1: Study set-up: Test material in ETR was created through a participatory design process and tested with 20 ETR users.

Figure 2 :
Figure 2: Smartphone test screens with design variant A (visible menu) embedded in functional homepage.

Figure 3 :
Figure 3: Smartphone test screens with design variant B (navigation overlay) embedded in functional homepage.

4. 2 RQ 2 :
Understanding and employment of digital design patterns 4.2.1 Visual knowledge of icons.Often, the only visual representation of a digital design pattern is a small icon, like the hamburger menu's horizontal bars on a mobile interface or the zoom indicator

Figure 4 :
Figure 4: The visual knowledge of icons that represent design patterns, when shown isolated (Questionnaire 1) and in a schematic drawing (Questionnaire 2).

Figure 5 :
Figure 5: The actual usage of the pattern 'hamburger' in the smartphone test.

Figure 6 :
Figure 6: The operation of digital design patterns in the smartphone test.

Figure 7 :
Figure 7: The screens of the PDF (left) and smartphone (right) difer in terms of cognitive load.The desktop screen displays more text and control bars which must be understood and managed.

Table 1 :
Tasks for both tests.The related screens of the smartphone test are shown in fgure 2 and 3.
Is the type size big enough?Can you enlarge?Zooming Is the two-fnger gesture for zoom used?(For PDF: is the navigation bar used?) 1 APlease go to entry summary.

Table 2 :
Age, lea reading level, access to the internet; and devices, applications, and usage.