Leveling the Playing Field: Temporal Video Segmentation for Individuals with ADHD in Computing Education

Synthesis: Pimenova, Begel and colleagues evaluate a post-hoc video processing intervention that segments instructional videos into single-instruction chunks with fixed pauses, reducing extraneous cognitive load for learners with ADHD. In a within-participants study (17 ADHD, 10 non-ADHD), the intervention improved everyone but had an equalizing effect: ADHD participants' errors and hesitations fell to parity with non-ADHD peers. This is strong evidence for Universal Design for Learning via automated content transformation — a task well-suited to AI pipelines. It extends I can't read your mind": A Study of Neurodivergent Computing Students' Experiences with Collaborative Active Learning and Inclusive Learning, complements Simulation work like LLM-Based Educational Simulation: Evaluating Temporal Student Persona Stability Across ADHD Profiles, informs video-based learning design in EduGage: Methods and Dataset for Sensor-Based Momentary Assessment of Engagement in Self-Guided Video Learning and Learning Design, within CS Education.

Key Findings

• A post-hoc video processing intervention segmented instructional videos into single-instruction chunks with fixed pauses to reduce extraneous cognitive load for learners with ADHD.
• In a within-participants study (17 ADHD, 10 non-ADHD), the segmentation improved outcomes for everyone.
• The intervention had an equalizing effect: ADHD participants' errors and hesitations fell to parity with their non-ADHD peers, not merely improving but closing the gap.
• The work is presented as evidence for Universal Design for Learning achieved through automated content transformation, a task well-suited to AI/Large Language Models (LLMs) pipelines.

What this means for practice

• Instructors. Split instructional videos at each single instruction and insert pauses rather than playing them straight through: on medium and hard Scratch tasks, participants with ADHD produced roughly 87% fewer errors on segmented versions of the same audiovisual content (β = −2.05, IRR = 0.128, p < .001).
• Instructors. Apply the treatment most aggressively to the hardest content, because effects grew with task difficulty: rate ratios reached 7.75 (errors) and 4.70 (hesitations) for the ADHD group, against 3.33 and 2.71 for non-ADHD peers.
• Instructors. Do not lean on re-watching as the accessibility strategy. Several participants described repeat-the-steps tasks as an implicit "test of memory" and asked for instruction to pause until the learner completes the corresponding action in the workspace.
• Designers. Ship segmentation as a post-hoc transformation of existing videos — nothing needs re-recording — with pauses held near 4 seconds, the duration a 13-participant pilot identified as long enough for mental rehearsal in Scratch's spatial-logical mapping while avoiding the engagement drops and mind-wandering seen at 6 seconds.

Limitations

• The sample is 27 participants (17 with ADHD, 10 without) recruited through campus flyers and non-computer-science faculty email at one university, all with no prior programming experience; the authors state that the sample and the Scratch block-coding environment may limit generalizability.
• Cognitive load was operationalized behaviorally only — errors and hesitations, where a hesitation is a 3-second pause or a verbal expression of confusion — with no physiological or dual-task measures (pupillometry, EEG, secondary-task interference) and no standardized working-memory assessment such as an n-back test.
• The segmentation-by-ADHD interaction terms were not statistically significant (errors p = .232; hesitations p = .242), and the authors state the study was underpowered to distinguish the groups statistically, so the equalizing claim rests on effect-size magnitude (ADHD d = 1.14 vs control d = 0.72).
• The 4-second pause is a fixed, pilot-derived parameter from a 13-participant pilot, and the segments were added manually in Final Cut Pro and agreed on by the authors, not generated automatically.

Citation

Veronica Pimenova, Chris Lee, Baramee Bhakdibhumi, Simon Chu, Andrew Begel (2026). Leveling the Playing Field: Temporal Video Segmentation for Individuals with ADHD in Computing Education.