1 of 12

Contextual Scaffolding and Self-Efficacy: Supporting Computer�Skill Development among Blind Learners in India��2026 Web Science & Digital Libraries Research Group Expo��Published in CHI ‘26: Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (ACM CHI - Barcelona, Spain)��

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

1

6/19/26

Authors : Akshay Kolgar Nayak, Yash Prakash, Sampath Jayarathna,

Hae-Na Lee, and Vikas Ashok

Department of Computer Science

Old Dominion University

Norfolk, Virginia, USA

@WebSciDL @AkshayKNayak7

Akshay K Nayak - Portfolio

2 of 12

Motivation and Context

India is home to ~70M visually impaired & ~4.95M blind individuals — the largest population globally
Less than 10% are formally employed (contrast: 44% in the US, 27% in the UK)
Government initiatives (Digital India, Skill India) and NGO-run computer training centers are where most BVI individuals first access computers and screen readers
Existing research on BVI computer education is almost entirely from the Global North, where accessibility resources and legal frameworks are more mature
Little is known about the in-situ teaching and learning challenges faced by trainers and BVI students in resource-constrained, multicultural contexts like India

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

2

6/19/26

“A persistent skills gap exists — BVI graduates remain confined to routine application-specific tasks”

@WebSciDL @AkshayKNayak7

So why India, and why this study? India has the world's largest visually impaired population — roughly 70 million people, including nearly 5 million who are blind. Yet less than 10% are formally employed. Compare that to 44% in the US and 27% in the UK — the gap is staggering.

To bridge this, government programs like Digital India and Skill India, along with NGO-run computer training centers, have become the primary places where BVI individuals first encounter desktop computers and screen readers.

But here's the problem: almost everything we know about inclusive computer education for BVI learners comes from the Global North — contexts with mature accessibility legislation, professional teacher training, and well-resourced infrastructure.

Very little research has examined what actually happens on the ground in resource-constrained, multilingual settings like India — how trainers teach, how students learn, and where the system breaks down.

3 of 12

Research Questions

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

3

6/19/26

@WebSciDL @AkshayKNayak7

RQ1: How do 'computer trainers' in India teach computer skills to BVI learners, including the curricula they use, teaching and assessment methods, and the challenges faced?

RQ2: How do the lived experiences of novice BVI learners shape their learning behaviors and self-efficacy as they navigate the pedagogical and technical challenges of resource-constrained computer training?

4 of 12

Methodology

Four-month contextual inquiry at two training centers in India (urban Bengaluru + rural Ranebennur)
94 BVI participants — trainees, computer trainers, and instructors
Three phases: Pre-study interviews → In-situ observations → Post-study interviews with graduates
Hybrid thematic analysis — inductive coding mapped to Bandura's self-efficacy and Vygotsky's ZPD

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

4

6/19/26

@WebSciDL @AkshayKNayak7

Illustration created by Gemini using the Nano Banana 2 model

We conducted a four-month contextual inquiry at two sites in India — an urban computer training center in Bengaluru focused on job-oriented IT certification, and a rural blind school in Ranebennur that follows the standard K–12 curriculum with basic computer instruction.

Across both sites, we worked with 94 BVI participants — trainees, computer trainers, and other instructors. The study had three phases: we began with pre-study interviews with newly enrolled students to understand their backgrounds and expectations. Then we spent two months observing at the start of the academic session and two months near the end — visiting twice a week, sitting in on classes for four to five hours, documenting how trainers taught and how students interacted with the material. Finally, we conducted post-study interviews with 23 recent graduates and 8 past trainees.

For analysis, we used hybrid thematic analysis — starting with inductive open coding grounded in participants' own language, then mapping emergent themes to Bandura's self-efficacy theory and Vygotsky's Zone of Proximal Development.

5 of 12

Findings: Misaligned Curricula & Unstructured Pedagogy (RQ1)

Curriculum followed KEONICS certification designed for sighted learners — SR training treated as a peripheral add-on, not certified or assessed
Trainers hired solely for SR proficiency, no formal pedagogy or CS credentials
Teaching relied on rote demonstration — trainer and trainee sharing earphones, repeating shortcut sequences
One-on-one demos were unscalable; students told to "come back when trainer was free"
Trainers avoided newer software versions: "We use Office 2007 only, because the new one is very complex"

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

5

6/19/26

@WebSciDL @AkshayKNayak7

"Screen readers cannot be certified by the government because they are not for general population. Screen readers are only for our personal use." — Trainer

6 of 12

Findings: Socio-Technical Factors & Self-Efficacy (RQ2)

Most trainees came from economically disadvantaged backgrounds — first meaningful computer access was at the training center
Late-blind learners received no transitional support — grouped with congenitally blind peers, expected to adapt immediately
Smartphone literacy did not transfer to desktop computer skills
Subjective norms: families celebrated any "respectful office job" but lacked knowledge of SR or accessible computing
Computer proficiency became a badge of pride — resisting societal perceptions of dependence

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

6

6/19/26

@WebSciDL @AkshayKNayak7

"Oh, even without vision, he's able to operate a computer — this recognition fueled their hopes." — Participant

Most trainees came from economically disadvantaged backgrounds — many held Below Poverty Line cards. For them, the training center was their first real encounter with a desktop computer.

Late-blind learners faced an added challenge: they received no transitional support. They were grouped with congenitally blind peers and expected to adapt to audio-based SR navigation immediately — despite having spent years relying on vision. Many had to unlearn the mouse-and-icons paradigm before they could even begin.

Importantly, being a smartphone user didn't help — that literacy simply did not transfer to desktop skills.

Families were supportive but lacked any knowledge of screen readers or accessible computing. They celebrated any 'respectful office job' without understanding what skills it actually required.

Yet despite all this, computer proficiency became a powerful motivator — a badge of pride. As one participant described, when others reacted with surprise that a blind person could operate a computer, that recognition fueled their hopes of building a career.

7 of 12

Findings: Learning Challenges & Behaviors (RQ2)

Compounded cognitive load: simultaneously learning keyboard layout, English terminology, SR shortcuts, multilingual typing, and mental models of software
Students lacked vocabulary to describe errors 🡪 self-learning resources were inaccessible (documentation too technical, assumed visual context)
Peer learning filled gaps but introduced confusion 🡪 different shortcuts for same task, inconsistent guidance
Skills remained confined to trained applications 🡪 no transferable navigation strategies
Learning stopped at graduation 🡪 without employment, students lost access to computers and skills eroded

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

7

6/19/26

@WebSciDL @AkshayKNayak7

"After the first few interview rejections, I lost confidence... I didn't know who to ask to learn more." — Participant

Learners faced a compounded cognitive load — they were simultaneously learning the QWERTY keyboard layout, English terminology, SR shortcuts, multilingual typing with Inscript mappings, and building mental models of software they'd never seen. All at once.

When they hit errors, they lacked the vocabulary to even describe what went wrong. Documentation was too technical, assumed visual context, and used terms like 'ARIA' or 'toggle' that meant nothing to them. Self-learning was effectively inaccessible.

Peers stepped in, but introduced confusion — students got different shortcut sequences for the same task and just went with whichever worked first. Skills also remained brittle and application-specific. A student proficient in MS Word avoided Google Docs because she 'didn't know how to use Docs' — no transferable navigation strategies were being built.

Most critically, learning stopped at graduation. Without employment, students returned home, lost access to computers, and skills eroded. As one participant said: 'After the first few interview rejections, I lost confidence — I didn't know who to ask to learn more.

8 of 12

Discussion — Inclusive Training in Resource-Constrained Contexts

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

8

6/19/26

@WebSciDL @AkshayKNayak7

Global North	India (Our Findings)
BVI-centered curricula (e.g., HKNC's TEAM) foreground accessibility	Ocularcentric, sighted certification exams (KEONICS); SR treated as an add-on
Trainers have pedagogy + special education credentials	Trainers hired solely for personal SR proficiency
Multi-stage, scaffolded AT exposure from K-12 onward	Compressed into one overwhelming year after high school
Multimodal methods (tangible, haptic, tactile)	One-dimensional audio-only SR interaction
Post-training rehabilitation pipeline sustains employment	Learning stops at graduation; self-efficacy collapses after job rejections

9 of 12

Discussion — Culturally Responsive Pedagogy for Higher Self-Efficacy

Current pedagogy imposes a uniform "blind" identity, discounting learners' diverse onset, language, and prior AT experience
Recommendation 1: Use native language analogies to introduce concepts first, then gradually transition to English SR commands — reducing extraneous cognitive load
Recommendation 2: Create assistive learner profiles grounded in onset of disability, prior AT familiarity, and linguistic background — enabling personalized scaffolding
Recommendation 3: Equip trainers with a pedagogical mirror — training-of-trainers program that builds conceptual knowledge of computers and SRs alongside teaching skills
K-12 curriculum should anticipate potential vision loss and introduce multi-sensory computer learning earlier

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

9

6/19/26

@WebSciDL @AkshayKNayak7

A core problem is that current pedagogy treats all learners as uniformly 'blind' — ignoring differences in onset, language, and prior AT experience.

We propose three recommendations. First, introduce new concepts using native language and familiar terminology before transitioning to English SR commands — reducing extraneous cognitive load so learners can focus on the intrinsic load of computer concepts.

Second, create assistive learner profiles grounded in onset of disability, prior AT familiarity, and linguistic background. A late-blind learner needs re-adoption scaffolds that are very different from what a congenitally blind learner needs. This enables personalized scaffolding rather than a one-size-fits-all approach.

Third, equip trainers through a training-of-trainers program — what we call a pedagogical mirror — that builds conceptual knowledge of computers and SRs alongside actual teaching skills.

Finally, K–12 curricula should anticipate potential vision loss and introduce multi-sensory computer learning earlier, not defer everything to a single post-school program.

10 of 12

Discussion — Designing EdTech for BVI-Inclusive Computer Learning

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

10

6/19/26

@WebSciDL @AkshayKNayak7

Current Gap	Potential Bridge
Overextended trainers can't provide stepwise, individualized instruction	AI-driven EdTech maintains learner profiles and adapts scaffolding in real-time
Students fear "breaking" the computer during exploration	Audio-based sandbox simulations for safe, consequence-free practice
SR output is in unfamiliar English; self-learning resources are inaccessible	Multilingual LLM coach that shadow-interprets SR output in native language
Trainers are treated as content conduits, not learners themselves	Trainer-facing tools with simulated novice BVI learners and pedagogical feedback
ATs function as black boxes to families and sighted peers	Design for interdependence — make learning needs visible to support networks
LLMs carry ocularcentric and ableist biases from training data	Participatory design with BVIs; audit for ableist language before deployment

Locally adaptable EdTech can fill critical gaps as ecosystems of inclusive training mature

Trainers are overextended and can't provide individualized instruction — AI-driven tools can maintain learner profiles and adapt scaffolding in real-time based on interaction traces like error patterns and response times.

Students feared 'breaking' the computer during exploration — audio-based sandbox simulations can provide a safe, consequence-free environment to practice commands without that anxiety.

SR output is in unfamiliar English and self-learning resources assume visual context — a multilingual LLM coach could shadow-interpret SR output in the learner's native language with contextual expansions.

Trainers themselves need support — trainer-facing tools with simulated novice BVI learners could let them rehearse lessons and receive pedagogical feedback.

And critically, any AI system deployed here must be audited for ableist and ocularcentric biases — LLMs trained on Global North data risk describing interfaces in purely spatial terms rather than structural ones. Participatory design with BVI users is essential before deployment.

11 of 12

Limitations & Future Work

Limitations:

Participants from a single region (Karnataka) — may not generalize across India's diversity
Primarily fully blind participants — excluded those with residual vision
Did not directly study sighted teachers in mainstream inclusive schools

Future Work:

Extend to multiple states and training centers
Include partial vision learners and mobile SR usage
Design and evaluate proposed AI-driven EdTech scaffolds with BVI learners and trainers

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

11

6/19/26

@WebSciDL @AkshayKNayak7

12 of 12

Closing Remarks

We present one of the first in-situ studies of how BVI learners acquire computer skills in resource-constrained, multicultural Global South contexts
The core problem is not just inaccessible tools but a deeper misalignment between Western-designed technologies/pedagogy and learners' situated realities
We advocate a shift from functional accessibility to culturally responsive computing pedagogy, supported by locally adaptable contextual scaffolds

Contextual Scaffolding and Self-Efficacy: Supporting Computer Skill Development among Blind Learners in India

12

6/19/26

@WebSciDL @AkshayKNayak7

Contact

Email: anaya001@odu.edu [Seeking Post-Doc or Full-Time Opportunities for 2027]

Akshay K Nayak - Portfolio

BlueSky: @akshayknayak.bsky.social

X: https://x.com/AkshayKNayak7