Project in Lieu of Thesis · committee draft
An accessibility-based approach to school connectedness for neurodivergent students
Project in Lieu of Thesis · committee draft
An accessibility-based approach to school connectedness for neurodivergent students
Neurodivergent students, such as those who are autistic, dyslexic, or have ADHD, are among those least likely to feel they belong at school. Belonging predicts mental health, academic achievement, and life outcomes well into adulthood. Most programs that promote belonging work through social routes such as peer attitudes and social skills. School for All takes a different route: accessibility. The program is built on the premise that a classroom designed around one way of learning forces neurodivergent students to spend energy overcoming the environment before they can learn, and that this hidden, avoidable effort accumulates into disengagement and, eventually, exclusion. Through lived-experience storytelling, assistive technology and study strategies, and stigma reduction, the program helps students participate and thrive now, while the longer work of universal design continues. The program has been implemented in all 10 compulsory schools in Hafnarfjörður, Iceland — reaching staff, students in grades 8–10, and families — and is now expanding to new municipalities and grade levels.
A powerful determinant of young people’s health is school connectedness, which describes the degree to which a student perceives that they belong in their school (CDC, 2024). Belonging is a fundamental human need (Baumeister & Leary, 1995), and in Maslow’s hierarchy it sits below esteem and self-actualization, meaning students must feel they belong before they can fully engage, perform, and grow (Maslow, 1943). A foundational review established belonging as a basic need the school community itself must meet (Osterman, 2000).
Youth who perceive high school connectedness have stronger academic outcomes, are more likely to graduate or pursue further education, engage in more positive health behaviors such as physical activity, and show better mental-health outcomes that persist into adulthood (Balfanz et al., 2024; CDC, 2024; Steiner et al., 2019). In contrast, feeling disconnected at school coincides with higher rates of emotional distress, suicidality, violence, and substance use, reducing adolescents’ quality of life across mental health, sexual health, and addiction, a pattern first documented in the foundational analysis of the U.S. National Longitudinal Study of Adolescent Health (NLSAH) in 1997 (CDC, 2023; CDC, 2024; Jones et al., 2022; Korpershoek et al., 2020; Raniti et al., 2022; Resnick et al., 1997).
School connectedness is a globally measured health outcome, and the picture the data show is concerning. International datasets such as the Programme for International Student Assessment (PISA), conducted by the Organisation for Economic Co-operation and Development (OECD) every three years in over 80 countries, and national datasets such as the U.S. Youth Risk Behavior Survey (YRBS), conducted by the Centers for Disease Control and Prevention (CDC), and Iceland’s national youth study (Íslenska æskulýðsrannsóknin) measure school connectedness at the population level. Across OECD countries in the 2022 PISA cycle, roughly one in four students did not agree that they belonged at school (OECD, 2023). In the United States, 45% of high school students in 2023 did not report feeling close to people at their school — the YRBS measure of school connectedness (CDC, 2024). In Iceland, one in five students reported not feeling they belonged at school in 2022 (OECD, 2023). Before the pandemic, in 2018, the number was even higher: one in four Icelandic students reported not feeling they belonged (OECD, 2019).
Factors that promote school connectedness in the general population have been extensively studied. A meta-analysis of 51 studies found that teacher support is among the strongest correlates of belonging at the secondary level (Allen et al., 2018). Additional factors include positive teacher–student relationships, perceived teacher fairness, parental emotional support, academic motivation, and a safe and fair school climate (Allen et al., 2023). Evidence from the NLSAH likewise links positive classroom management climates, tolerant discipline policies, participation in extracurricular activities, and smaller school size to higher connectedness (McNeely et al., 2002). However, these factors have largely been studied in the general student population, and special consideration is needed for student populations who face additional barriers in school.
Neurodivergent students, such as those who are autistic, dyslexic, have ADHD, or other neurodevelopmental differences, are among those at greatest risk of low school connectedness. Together, these conditions are estimated to affect 15–20% of school-aged youth (Littlefair et al., 2024; Yang et al., 2022). The proportion is likely higher. Autism is globally underidentified, particularly in girls (Zeidan et al., 2022; Lockwood Estrin et al., 2021), and adult ADHD is widely underdiagnosed (Song et al., 2021). Population-based Icelandic data confirm these patterns at the national level, documenting sex differences in identification and high rates of co-occurring psychiatric conditions among autistic and ADHD young people (Sigurdardottir et al., 2025). Additionally, diagnosis can be hard to access, particularly for low-income families, and waiting lists can be long (Aylward et al., 2021; Odegard et al., 2020; O'Nions et al., 2023). In Iceland, waiting lists can be up to three years for assessment or support (Biðlistinn, 2025). Consequently, a substantial group of students’ neurodivergence has not been — and may never be — formally identified, leaving them without formal support while facing the same barriers.
Research consistently shows that neurodivergent students report lower connectedness than their neurotypical peers (Tsou et al., 2025; Boshoff et al., 2025; Wilmot et al., 2024). The evidence, though limited, is consistent. An Icelandic study of 480 schoolchildren first compared students with and without chronic health conditions and found that those with chronic conditions reported lower school connectedness; within that group, students with learning disabilities and mental-health conditions were at the greatest risk (Svavarsdóttir, 2008). Adolescents with learning disabilities were found to be at twice the risk of emotional distress, and females were at twice the risk of suicide attempt, with connectedness to parents and school being the strongest protective factors against these outcomes (Svetaz et al., 2000). This risk also appears once avoidance has progressed to chronic non-attendance, a pattern studied in its own right in the school-attendance literature (Kearney & Graczyk, 2020). Autistic students miss roughly a fifth of school days, driven by school factors rather than unwillingness (Totsika et al., 2020), and in one large UK study, 92% of children in school distress were neurodivergent (Connolly et al., 2023). A student who avoids or is excluded from school is not a student who feels they belong.
Despite neurodivergent students’ elevated risk of low connectedness, data on this population remain scarce in national data collections. PISA, the YRBS, and Iceland’s national youth study all examine factors such as gender and socioeconomic status, but not disability or neurotype (CDC, 2024). This gap is well documented: International assessments such as PISA systematically exclude many students with disabilities through their sampling rules (Schuelka, 2013) and do not disaggregate special educational needs the way they do gender and income (LeRoy et al., 2019), and a peer-reviewed analysis confirms that the YRBS has no disability item (Lutz et al., 2023). Consequently, the proportion of disconnected students who are neurodivergent is unknown — the students most at risk are statistically invisible.
| Survey measures… | Gender | Socioeconomic status | Disability / neurotype |
|---|---|---|---|
| PISA (OECD) — 80+ countries | ✓ | ✓ | ✕ |
| Youth Risk Behavior Survey (US) | ✓ | ✓ | ✕ |
| Íslenska æskulýðsrannsóknin (Iceland) | ✓ | ✓ | ✕ |
The rare exceptions confirm the pattern. When a U.S. survey cycle carried a disability item, students with disabilities showed higher rates of sadness, suicidality, substance use, and victimization (Everett Jones & Lollar, 2008); when Ireland’s national cohort measured special educational needs, those children were significantly less likely to like school (McCoy & Banks, 2012); and in the United Kingdom, where a regional survey links belonging to special-educational-needs status, students with special educational needs report lower school belonging and markedly higher disability-based discrimination (BeeWell Programme, 2025). No national instrument yet captures this systematically.
The problem is global; Iceland is no exception, and it is the context in which this work is grounded. Iceland is a nation of approximately 390,000 people with a ten-year compulsory school system (grunnskóli). National policy embraces inclusive education (skóli án aðgreiningar), yet an external audit found major implementation gaps, including no shared understanding of inclusion in practice (European Agency for Special Needs and Inclusive Education, 2017), and Icelandic teachers describe inclusion as an added burden they were not prepared for (Gunnþórsdóttir & Jóhannesson, 2014). The global and Icelandic pictures point to the same place: the environment.
Many factors affect belonging for neurodivergent students, including bullying, difficulty connecting with peers, different communication styles, and ableist attitudes, all of which further elevate their risk. Crucially, differences in communication are a difference, not a deficit, a point increasingly made by autistic-led research, which locates communication breakdown between neurotypes rather than within autistic people (Milton, 2012; Marocchini, 2023; Crompton et al., 2020). These factors matter, and other approaches address them (Hodges et al., 2022; Kasari et al., 2012). This work does not claim to be the whole answer. Instead, it examines one specific, underexamined, and modifiable influence on school connectedness for neurodivergent students: inaccessibility. Based on that lens, I propose a conceptual framework that explains how inaccessible educational environments contribute to exclusion and, through it, to the erosion of school connectedness for neurodivergent students. I propose that belonging does not collapse all at once. It erodes — one avoidable barrier at a time — until a student stops showing up. Part 2 explains the development of this framework; the program in Part 3 is built on it.
I present this as a conceptual framework: a synthesis of established constructs from cognitive load theory, the social model of disability, and Universal Design for Learning into a single explanatory pathway, the Belonging Chain. It follows the tradition of pathway frameworks such as the minority stress model (Meyer, 2003), which explains how chronic stress arising from stigma and prejudice harms the health of minority groups, and the stress process model (Pearlin et al., 1981), which traces how conditions in a person’s social environment become stressors that wear health down over time. Such frameworks trace a distal environmental cause through an internal mediating state to an outcome (Jabareen, 2009; Imenda, 2014). The integration of these previously unconnected literatures into one pathway is itself the theoretical contribution (Whetten, 1989).
The factors that build school connectedness are well studied in the general student population (Allen et al., 2018), and existing belonging interventions for neurodivergent students work largely through social routes, namely peer attitudes and social skills (Hodges et al., 2022; Kasari et al., 2012). The leading socio-ecological framework of school belonging (Allen et al., 2016) maps the multi-level factors that foster belonging, but it does not model the mechanism by which environmental inaccessibility erodes belonging for neurodivergent students specifically. Cognitive load theory, the social model of disability, and UDL each illuminate part of why neurodivergent students disengage, but, to my knowledge, no existing framework connects them into a single pathway from environmental inaccessibility to belonging erosion. It has also been noted in the field that social-skills approaches risk locating the “fix” in the child rather than the environment (Bottema-Beutel et al., 2018). This framework takes the environmental route.
Four lenses each explain a link in the chain. Universal Design for Learning (Rose & Meyer, 2002) identifies the fault: a one-size-fits-all classroom (Link 1). Cognitive load theory (Sweller, 1988; Paas & van Merriënboer, 2020) supplies the mechanism: the extraneous effort inaccessible instruction imposes (Links 2–3). Stigma and stereotype-threat research (Haft et al., 2023; Shifrer, 2013) explains the turn inward: identity erosion (Link 4). And hidden-exclusion and disability-studies research explains the end of the chain (Links 5–6): exclusion is often informal and invisible (internal exclusion, part-time timetables, off-rolling, remaining on the register but not in class — forms documented most extensively in the UK), so official figures undercount it, and disabled and neurodivergent students are heavily overrepresented in these hidden forms (Power & Taylor, 2020; McCluskey et al., 2019; Connolly et al., 2023).
Three further lenses hold up the whole chain rather than any single link. The social model of disability (Oliver, 2013) and, complementing it, an occupational-justice perspective (Townsend & Wilcock, 2004) both locate the problem in the environment and the person–environment fit rather than in the student. The ideal response is therefore a learning environment that is inherently flexible, in line with UDL, so that fewer individualized accommodations — which are often burdensome to obtain — are needed in the first place. Individualized supports and assistive technology bridge the gap where universal design does not yet reach. Meyer’s (2003) minority stress model supplies the pathway’s form, and Maslow (1943) and the school belonging literature (Goodenow, 1993; Allen et al., 2018) frame the outcome.
1. Structural inequity. Traditional classrooms are designed around how neurotypical students learn, typically offering one means of taking in information and one means of demonstrating it. This under-implementation of Universal Design for Learning disadvantages students who learn differently from the start (Rose & Meyer, 2002; Bjartmarsdóttir & Newbutt, 2025).
2. Learning friction. Learning is never effortless, and it should not be. Learning friction is the cumulative cognitive load a student carries as the three kinds of load interact: intrinsic (how demanding the material itself is), extraneous (load added by how it is presented), and germane (the productive effort of actually learning) (Sweller, 1988; Paas & van Merriënboer, 2020). Some friction is natural, necessary, and even desirable, because germane effort is where learning happens and a degree of challenge deepens it (Bjork & Bjork, 2011). The problem is not friction itself but its total level. For a neurodivergent student in a classroom not designed in line with UDL, extraneous load rises disproportionately. Stacked on top of intrinsic and germane load, the cumulative friction can exceed the student’s working-memory capacity, which is already more heavily taxed for these learners (Martinussen et al., 2005; Swanson et al., 2009; Le Cunff et al., 2024). Sustained, this overwhelm blocks further learning and, over time, tells the student they are “not smart,” when the real problem is the load the environment added.
Picture two students studying one hour for the same exam. The neurotypical student opens the textbook and spends most of the hour learning. The dyslexic student opens the same textbook, but half of their mental energy goes to decoding the words rather than learning the content. In the same hour, the dyslexic student learns less and scores lower — through no fault of their own.
3. Cognitive depletion. As used here, cognitive depletion is the downstream cost of disproportionate friction: the student’s finite working-memory resources are spent on access before and during learning, so they enter overwhelm sooner, cover less material in the same study time, and — because learning took longer — arrive at the moment of demonstrating what they learned with their resources already drained. The result is worse performance despite equal or greater effort (Sweller, 1988; Paas & van Merriënboer, 2020).
4. Identity erosion. Studying as hard as one’s peers yet performing worse conveys the message that the problem is internal, that one is “not cut out” for school, eroding academic self-concept, self-efficacy, and belonging (Haft et al., 2023; Shifrer, 2013; Ragnarsdóttir, 2023).
5. Behavioral disengagement. Eroded belonging predicts withdrawal, and it does not happen overnight. Disengagement begins gradually — participating less, asking for less help, trying less — before it externalizes into visible behavior, a cycle described in the participation–identification model (Finn, 1989; Korpershoek et al., 2020; Quin, 2017).
6. Exclusion. As the disengagement becomes more pronounced, it hardens into exclusion — internal forms, in which a student remains enrolled but is withdrawn or removed from ordinary classroom life, and external forms, such as school avoidance and formal removal. Neurodivergent students are markedly overrepresented in both (Totsika et al., 2020; Connolly et al., 2023; Bond et al., 2007), and in England, students with special educational needs are several times more likely to be formally excluded than their peers (Timpson, 2019; Graham et al., 2019).
Note. Disengagement can take two forms: flight (avoidance and exclusion) or fight (overcompensation through overwork, at the cost of health). The “fight” path connects to a pattern of neurodivergent burnout into adulthood (Raymaker et al., 2020), developed as a separate extension of this framework in future work.
School connectedness / belonging — the degree to which a student perceives that they belong at school (CDC, 2024).
Structural inequity — the disadvantage produced when classrooms offer a single means of representation, engagement, and expression (under-implemented UDL).
Learning friction (a term introduced in this framework) — as used here, the cumulative cognitive load a learner carries as intrinsic, extraneous, and germane load interact. Some friction is natural and necessary, but a disproportionate rise in extraneous load, from an environment not designed for the learner, can push the total past working-memory capacity, causing overwhelm that inhibits learning, distinct from desirable difficulty (Bjork & Bjork, 2011).
Cognitive depletion — the exhaustion of finite working-memory resources before learning occurs.
Identity erosion — the decline in academic self-concept, self-efficacy, and belonging that follows repeated underperformance despite effort.
Behavioral disengagement — withdrawal from participation.
Exclusion — school avoidance, formal removal, or the hidden forms of exclusion in which a student remains enrolled but disconnected.
Hidden exclusion — practices that remove a student from mainstream classroom life while they remain on the school register, such as internal exclusion, part-time timetables, off-rolling, and informal removals, so that official exclusion figures undercount the true rate (Power & Taylor, 2020).
This framework applies to neurodivergent students (autistic, ADHD, dyslexic, and related learners) in conventional compulsory and secondary education environments not designed in line with UDL. It is not claimed to be the sole determinant of belonging, but rather provides insight into a previously unexplored factor of belonging for neurodivergent students. Bullying, peer connection, communication differences (Marocchini, 2023), and ableist attitudes continue to be prominent factors that impact belonging at school for these students. The framework is theoretically derived. Several links are correlational or reciprocal rather than strictly one-directional (notably Links 3–5), and the pathway describes elevated risk across a population rather than a fixed trajectory for any individual. Finally, this framework and the program built on it sit alongside, rather than in place of, other approaches: neurodiversity-education curricula such as LEANS build peer understanding and attitudes (Alcorn et al., 2024), and social-focused interventions build peer connection (Hodges et al., 2022; Kasari et al., 2012). The accessibility route taken here has not yet been tested against, or in combination with, these approaches.
Because the framework identifies learning friction as the upstream, modifiable factor, it is the program’s most direct point of intervention. Assistive technology (tools that reduce extraneous cognitive load and thereby increase equitable access to educational opportunities) and study strategies reduce learning friction by lowering the extraneous portion of the load (reformatting material, scaffolding assignments, and using AI to fill quick knowledge gaps) so that more of a student’s capacity is left for learning. When a dyslexic student listens to a textbook with text-to-speech instead of decoding it, the energy otherwise lost to reading is available for learning (Perelmutter et al., 2017; Bjartmarsdóttir & Newbutt, 2025). However, tools alone are insufficient. Students often avoid assistive technology for fear of appearing different (Vaccarella et al., 2025; Parette & Scherer, 2004), so reducing stigma is essential.
The framework showed how the environment erodes a student’s academic self-efficacy along the chain; the program is built to rebuild it, primarily through credible role models. Academic self-efficacy is thus the bridge between framework and program: the capacity the chain erodes, and the capacity the program is built to restore.
The program is grounded in three frameworks working at multiple levels: Social Cognitive Theory (SCT), the Social Ecological Model, and the social model of disability. SCT’s key constructs, including self-efficacy, observational learning, outcome expectations, reinforcement, social and normative influences, environmental facilitators and barriers, and reciprocal determinism, provide a foundation for understanding the processes that shape school connectedness (Bandura, 1986; National Cancer Institute, 2005). The Social Ecological Model situates the student within nested levels (individual, relationships, school, family, and community), and the program intervenes across them rather than at the level of the student alone (McLeroy et al., 1988). The social model of disability frames the barrier as residing in an inaccessible environment, not in the student (Oliver, 2013). Together these constitute the program’s theory of change, the logic of how its activities produce belonging.
The program’s overarching aim is to strengthen school connectedness for neurodivergent students, the primary outcome evaluated in Section 3.6. Four goals serve this aim, ordered by the program’s causal logic: lower learning friction first, rebuild what it eroded, and then shift the culture and environment that produced it. Measurable objectives are developed from these goals, following the CDC’s framework for program evaluation in public health, in which clearly stated goals and measurable objectives form the foundation of evaluation design (CDC, 2011).
Goal 1 — Reduce learning friction. Equip neurodivergent students with assistive technology and study strategies that lower the extraneous load of an environment not designed for them, so that more of their capacity is available for learning.
Goal 2 — Rebuild academic self-efficacy and hope. Restore students’ belief in their academic potential through credible role models and lived-experience evidence that their path is not fixed, laying the foundation for self-advocacy.
Goal 3 — Reposition neurodivergence and its tools. Reduce stigma among students, staff, and families by reframing assistive technology as a “study smarter, not harder” strategy that strong students use, and neurodivergent traits as strengths.
Goal 4 — Equip staff to lower barriers. Increase staff knowledge, confidence, and practical use of inclusive, neurodiversity-affirming strategies and assistive technology.
Goals 1 and 2 work at the level of the student. Goals 3 and 4 work at the level of the culture and environment around the student.
The aim-level outcome — school belonging — sits above all four goals and is measured with the PSSM plus the retrospective belonging item (Section 3.6).
The facilitator’s own story is used deliberately because evidence shows that narrative increases engagement and relevance and functions as indirect contact that shifts stigma-related attitudes (Nagarkar et al., 2026; Thornicroft et al., 2016; Corrigan et al., 2012). Two public-health principles guide this choice: the messenger must be credible and relatable to the audience, and while data provide context and credibility, it is stories that compel action (Wen, 2021). The story does specific work:
Repositioning perceptions. The through-line of the program is repositioning, changing how tools and traits are seen. Assistive technology is repositioned from “something only struggling students use” to a “study smarter, not harder” strategy that high-achieving students use too, and neurodivergent traits are repositioned from deficits to strengths within a neurodiversity-affirming, UDL frame (Pellicano & den Houting, 2022; Marocchini, 2023). This reframing lowers the stigma barrier to uptake and supports self-compassion (Corrigan et al., 2012; Thornicroft et al., 2016).
Hope and inspiration. The facilitator shares her own trajectory: a student who struggled in the Icelandic education system with multiple learning disabilities, and who, after a period of physical disability and hospitalization, has fully recovered and now thrives in academia. For students who, as she once did, are beginning to lose hope, this offers tangible evidence that their path is not fixed (Stout et al., 2011; Marx & Roman, 2002).
The role-model effect. The facilitator serves as a credible same-group role model that neurodivergent students can identify with, a mechanism that research links to self-efficacy, aspiration, and belonging for members of negatively stereotyped groups (Stout et al., 2011; Marx & Roman, 2002). This ingroup-role-model evidence is strongest for women in STEM (Science, Technology, Engineering, and Mathematics); here it is applied to neurodivergent students as an extension of the same mechanism.
The curb-cut effect. Every lecture connects the tools to the curb-cut effect, the principle that designs created for those who need them most benefit everyone, normalizing the tools for everyone (Blackwell, 2017; Hamraie, 2017; Szumski et al., 2017).
This design choice also enacts the disability-rights principle of “nothing about us, without us”: a program for neurodivergent students is designed, facilitated, and researched by a neurodivergent person. The lived experience is not an anecdote attached to the program — it is the source of the program’s credibility and a working part of its method.
Alongside the story, the program’s second instrument is explicit naming and describing of the conceptual framework and its constructs to students, staff, and families. Each audience receives a version adapted to its context and assumed health and research literacy, meeting learners where they are to ensure understanding.
The first concept named is hidden exclusion. Where the framework uses hidden-exclusion research to explain the end of the chain (Power & Taylor, 2020), the program turns it into a teaching point. Exclusion is not only formal removal but the everyday, indirect messages that tell a student they do not belong. A building with an “everyone welcome” sign but no ramp still tells a disabled person they were not expected to come. This example is used in the program to visualize the hard-to-visualize hidden exclusion that neurodivergent students face.
The second concept named is learning friction. In the staff and family sessions, the facilitator names cognitive load theory directly and walks through the Belonging Chain, showing how an environment not built for a student converts effort into overwhelm, and overwhelm into disengagement. In the student session, the facilitator translates this to learning. Repeatedly setting a student tasks they have neither the tools, the skills, nor the time to complete, in an environment that was not built for how they learn, sends the indirect message that something is wrong with the student rather than with the learning design. Learning design matters, because not everyone learns the same way. Some learn best by seeing, others by writing, listening, or doing. Each student has their own way, and finding it, while staying aware of learning friction, is part of learning how you learn best.
The third concept named is assistive technology. In all sessions, assistive technology is defined broadly: any tool that reduces friction between a learner and their environment — glasses, headphones, text-to-speech, timers, organizational tools, anything that helps. Within this definition, artificial intelligence is introduced explicitly as an emerging category of assistive technology and described to students, staff, and families as a tool that, used with guidance, can increase access to information, reading, writing, and organization. Naming AI directly gives all three audiences a shared, non-stigmatizing language for a technology students are already encountering.
In student sessions, this is made concrete through a study workflow grounded in evidence-based learning strategies: students pair AI with their own course materials to generate active-recall quizzes, use the results to map strengths and gaps in the knowledge required, direct their study time toward the identified gaps, and then talk through difficult concepts with AI as a form of self-explanation. Each step reflects established findings — that active learning outperforms passive instruction in K–12 settings (Tutal & Yazar, 2023), that retrieval practice is among the highest-utility learning techniques (Dunlosky et al., 2013), and that prompted self-explanation produces meaningful learning gains while helping students identify what they do not yet know (Bisra et al., 2018). AI is likewise presented as a scaffolding tool that can break demanding assignments into manageable steps — a new application of one of the best-established instructional principles (D. Wood et al., 1976; van de Pol et al., 2010). This enables older students to work more independently, helps parents support learning in subjects they have not mastered themselves, and lets teachers prepare scaffolded materials for a struggling learner in seconds rather than hours, freeing time for other students. The sessions also introduce complementary strategies for focus and executive functioning. The first is structured work–break intervals, such as the Pomodoro method, consistent with research on effort regulation and attention restoration. The second is body doubling — working alongside another person — an emerging, community-identified strategy with early empirical support among neurodivergent people (Eagle et al., 2024). Framed this way, AI and its companion strategies reduce friction not by doing the thinking for the student, but by making proven strategies easier to execute.
The sessions also emphasize that these tools help all students: text-based telephone communication was first developed for deaf people (the TeleTYpewriter), and texting is now part of everyday life for everyone. Every session therefore closes by naming the curb-cut effect — the principle that designs created for those who need them most end up benefiting everyone (Blackwell, 2017) — connecting the tools back to the program’s core repositioning message.
Books in the format that fits the reader
Listen to any text instead of decoding it
Fonts, spacing, and overlays that ease reading
Write by speaking
Catch and fix writing errors as you go
Translation and language-learning support
Step-by-step solvers and visual math tools
Planning, organizing, and time management
Tools that support calm and self-regulation
Blockers and settings that protect focus
Fill knowledge gaps and scaffold tasks
Adjusting light, sound, and space to the learner
The program consists of five core components, structured around the Social Ecological Model: four session types that reach the nested levels of the student’s environment, and educational materials that sustain the work after delivery. All components are delivered universally to whole schools rather than based on diagnosis; the program therefore also reaches the substantial group of students whose neurodivergence has not been — and may never be — formally identified (Section 1.4). The components are the fixed core of the program, while their dosage (session length, number of sessions, and audience size) is adapted to the needs of each community. The model is designed for grades 5–12. It was first implemented in Hafnarfjörður in grades 8–10, and a successful grades 5–7 pilot now informs its expansion to grades 5–10 in a second municipality.
Staff session (school level; Goals 3–4) — 60 minutes for all school staff, including administrators, focused on reducing stigma, showing neurodivergent potential, and giving practical assistive technology and inclusive strategies for immediate classroom use.
Student sessions (individual level; Goals 1–2) — 30 minutes, one per grade level, plus targeted sessions for special-education groups and online sessions for students in school exclusion, all built around the lived-experience story: assistive technology, study strategies, strengths, self-understanding, and hope. The online sessions extend the program to students at the far end of the Belonging Chain.
Family session (family level; Goal 3) — for parents and other family members, in person and online, sharing the same repositioning story and assistive technology for use at home, supporting belonging beyond school.
Follow-up cohort (school and institutional level; Goal 4) — two 60-minute sessions with teaching consultants, engaged teachers, and administrators from the staff session, focused on reducing structural barriers, applying assistive technology school-wide, and laying the groundwork for institutional change. Who attends this cohort shapes how far the program’s effects reach.
Educational materials (all levels) — Notion templates, assistive-technology tutorial videos (reading and writing support, organization tools), and slides for ongoing in-house training, designed to extend the program’s effects beyond the single contact.
The tools presented are deliberately many and varied: not every tool will work for every student, or be feasible for every student to use, but the breadth ensures that everyone can find something that works for them — across all areas of learning difficulty, addressed in different ways. The assistive technology taught across the sessions spans twelve categories (see Figure 12; Appendix A lists example tools and supporting evidence by category): digital and audio books; text-to-speech; reading support; speech-to-text; grammar support; language; math; executive functioning and time management; emotional regulation; attention tools; AI; and sensory-friendly learning environments. Several categories carry direct evidence: text-to-speech improves reading comprehension for students with reading disabilities (S. G. Wood et al., 2018), and the gamified approaches used in the math/STEM and language tools show positive effects on cognitive, motivational, and behavioral learning outcomes (Sailer & Homner, 2020). The math and language categories are best understood not as distinct accessibility mechanisms but as UDL’s multiple means of representation, engagement, and expression operationalized in tool form (CAST, 2024). Sensory-friendly learning environments include low-cost regulation supports — earplugs, noise-reducing headphones, fidget tools, adjustable lighting, and weighted or compression items. These are offered as regulation options rather than as evidence-based interventions: their purpose is to reduce sensory overwhelm and support recovery and regulation during and after cognitively demanding tasks, so the student can return to a state in which learning is possible. Evidence here is emerging and mixed — fidget devices, for example, have improved on-task classroom behavior for students with ADHD (Aspiranti & Hulac, 2022), while reviews of sensory-based tools more broadly urge caution — which is why the program presents them as options to try, not prescriptions.
A multi-level design spanning the student's whole environment — staff, students, support staff and families — grounded in SCT and the Social-Ecological Model.
Beyond the classroom, the program aims to shift norms and cultural views of disability across these levels. This is a long-term ambition that no single evaluation can capture, but it is the ambition the socio-ecological lens makes explicit. This multi-level shape is consistent with evidence that the most effective school mental-health programs target everyone, aiming to change culture rather than treat individuals (Szeto et al., 2024), and that whole-school approaches with family components outperform single-component ones (Goldberg et al., 2019; Allen et al., 2022).
As an initial, feasibility-focused evaluation of the program’s goals, the primary outcome, school belonging, is measured with the Psychological Sense of School Membership (PSSM) scale (Goodenow, 1993), completed by students after the program via Microsoft Forms on their personalized school devices, in Icelandic or English depending on the student’s language of instruction, together with a short set of researcher-designed items delivered in the same language. The PSSM is currently being translated into Icelandic by the researcher and back-translated into English by a bilingual reviewer with subject-matter expertise and prior back-translation experience, who had no access to the original English version; a second bilingual reviewer with prior experience in Icelandic-context translation is comparing the back-translated version to the original for semantic equivalence, with discrepancies reconciled through discussion and revision, following established back-translation procedures (Brislin, 1970). To the author’s knowledge, this is the first Icelandic adaptation of the PSSM — an instrument contribution in itself — with known method effects from negatively worded items addressed in scoring (Ye & Wallace, 2014). Researcher-designed items were written to match the literacy level of the intended audience and incorporated images of AT tools and categories to reduce reading demand, consistent with the study’s UDL framework. Keeping the survey brief fits the single-session format. Because change is the target, the items use a retrospective post format, which reduces the response-shift bias that can distort conventional pre/post self-reports (Howard, 1980).
In line with the GOMS table, the researcher-designed items capture, in order: perceived inspiration and hope (Goal 2); retrospective change in school belonging (the program’s aim); willingness or intention to use at least one tool or strategy from the program (Goals 1 and 3); perception of prior difficulty with school or learning (an indirect indicator); and an optional open-ended item on which tool the student is most excited to try. Exact item wording is finalized during ethics review.
The final item is indirect. Rather than asking for a diagnosis, it flags students who have experienced difficulty in school or learning. This allows a test of the hypothesis that students who have struggled, those the program is designed to reach, benefit more from it. As the study concerns students’ experience of their own learning, it would be submitted to the Icelandic National Bioethics Committee (Vísindasiðanefnd) to confirm the required level of oversight before any data collection. After a conversation with the National Bioethics Committee, they have expressed that this data collection does not need a formal review as the data collected is not direct health data and is collected anonymously. This initial evaluation focuses on the aim and the student-level goals; the culture- and staff-level goals will be evaluated in later phases.
The program is deliberately limited in scope. It does not attempt to reform, on its own, the structural inequity of an education system not yet designed in line with UDL. Embedding UDL system-wide is the work of many researchers over many years. Encouragingly, the field is moving in this direction: the most recent UDL Guidelines explicitly foreground identity and belonging as design goals (CAST, 2024). Instead, the program helps neurodivergent students survive and, ideally, thrive within the environment they are in now, while that longer systemic change proceeds. Introducing all staff to UDL plants a seed for it. As a short, single-contact intervention, its measurable effects are correspondingly modest and near-term, and the evaluation above is an initial, feasibility-oriented assessment rather than a definitive test. The framework it rests on is theoretically derived and awaits empirical study.
School for All was first delivered across all 10 compulsory schools in Hafnarfjörður, Iceland, over one school semester (February–April 2026), reaching staff, students in grades 8–10, and families. The implementation began with a request: after a 2025 television interview, a parent, Björg Sæmundardóttir, connected the program’s developer with the town’s mayor, who asked for a program for all ten schools (V. Víðisson, personal communication, October 8, 2025). At a December 2025 stakeholder session, the municipality proposed adding a session for parents — initially a workaround for privacy rules that limit assistive technology on school devices, and subsequently a permanent component, since the home is where access is least restricted. Schools were run in three groups of roughly a month each, so sessions could be refined from feedback as delivery progressed.
Unsolicited verbal feedback from staff, students, and families throughout delivery informed real-time adaptation. Several key changes resulted. The student session was reframed after staff reported students refused tools to avoid looking “different” — disability language was dropped (except the facilitator’s own story), and tools were marketed instead as a universal “study smarter, not harder” strategy. Dedicated sessions for special-education groups were added after students in special-education classes and námsver (study labs) were found to be invisible in the grade-level plan; this is now a standard component. Delivery shifted to one grade level at a time for feasibility, and the student session was tightened to reduce cognitive load. Tool-name handouts were added at students’ request. Finally, the staff follow-up was redesigned into a shorter, hands-on format while the municipality cleared tool barriers.
The first cycle showed that a municipality-wide, multi-level program is feasible: all ten schools were reached in a single semester. Demand is real, as the program was requested by the municipality itself and is now being adapted for others. The delivery model also proved able to correct itself; acting on unsolicited feedback turned a first draft into a refined, replicable model. The most striking lesson came from the program’s own framework — students in special-education settings were initially left out because schools did not report them, hidden exclusion operating quietly inside an inclusion program (Power & Taylor, 2020). The episode strengthens the program’s core claim: exclusion operates invisibly unless actively countered.
The first cycle was not formally evaluated, and no research data were collected, as the work preceded ethics review. Refinements were instead informed by unsolicited verbal feedback from staff, students, and families, noted as it arrived — the engine of the adaptations above. Formal evaluation using the Icelandic PSSM described in Section 3.6 is planned for future cycles, following the required ethics approval, with waitlist or stepped-wedge designs under consideration so that later schools can serve as comparison groups (Brown & Lilford, 2006).