How We Will Measure Success While a Child Uses Hero's Embrace
Step 4 of the walkthrough shows what a child experiences inside a Skill Quest. This page answers the question a reviewer asks next: how do we know it's working? It walks through the measurement architecture behind the Child Experience — how activity and progress data is captured, how the story itself is designed to function as the data-collection instrument, and how growth in the platform's three functional domains — Adaptive, Executive, and Social — is practiced and tracked session by session, with metacognition built in as the underlying process rather than a separate score.
Hero's Embrace is not claiming clinical efficacy today, and this page does not either. The Phase I feasibility pilot (N=4-6, Students/Families, 12 weeks) is designed to determine whether children engage meaningfully with the platform, whether the Adaptive Learning Orchestrator (ALO) keeps each child in a healthy zone of challenge, and whether caregivers and educators can act on what the platform reports back to them. It is explicitly not powered to test whether the platform improves outcomes — that is the question a Phase II randomized controlled trial (N=300) is designed to answer, using the feasibility data and refined protocol Phase I produces.
What follows is the measurement design behind that feasibility question: the mechanisms already specified for capturing activity and progress data during ordinary play, not a set of clinical claims.
Every measurement touchpoint below already exists in the Placement Guidelines, the SparkProfile spec, and the NIH SBIR Research Strategy — this section connects them into one pipeline, from a child's first check-in to the data a Phase I analysis would actually report.
Three things travel through that pipeline, all specified in the SBIR Research Strategy's Aim 2:
What the ALO reads during play
- Performance signals — which choice a child made, how long they took, whether they used a tool or hint.
- Behavioral signals — task initiation delay, whether a child restarted or abandoned a beat.
- Affective signals — frustration or disengagement indicators, monitored against a pre-specified threshold.
- Longitudinal trajectory — how a child's pattern in a domain shifts across sessions, not any single session in isolation.
What that produces, downstream
- Domain-level trend lines for Adaptive, Executive, and Social Function — not a single score, a trajectory.
- A re-administered check-in at natural milestones (e.g., a tier change, or the Month-6 interim point in the SBIR timeline) to see if scores and any parent/teacher discordance have shifted.
- A Transfer Moment Debrief — the same Reflect moment, in two forms: private to the child, and a short plain-language note to the caregiver/educator.
- An Explainable AI Dashboard entry — what data was used, what the system adjusted, and why, in language a non-technical caregiver can read.
Hero's Embrace does not bolt an assessment onto a game. The Build → Solve → Reflect loop is itself the measurement mechanism, and it is not an original design choice made in a vacuum — it is based on a structured Focus & Plan → Act → Reflect approach used in evidence-informed interventions for children who struggle with attention, self-regulation, and adaptive functioning. (A formal named citation for this approach is pending confirmation with the outside researcher whose program it draws from — not yet finalized, so it isn't named here.)
Build → Focus & Plan
- The Companion sets up a concrete stake — a checklist, a bridge, a disagreement — never an abstract prompt.
- Body/state-awareness questions come first ("what does 'too loud' actually feel like in your body?") — noticing before fixing.
Solve → Act
- The child chooses from a bounded set of developmentally scaffolded options — not an open-ended free-response, which the target population's language/executive profile makes a poor measurement surface.
- Which option, how long to decide, and whether a tool or hint was used is the behavioral data point.
Reflect → Return with Insight
- The Companion reflects back honestly on whatever was chosen — never grading the choice, always naming something true.
- This single moment has two audiences at once: a private recap for the child, and the Transfer Moment Debrief sent to the caregiver/team. They are two views of the same honest moment, not two separate write-ups.
The validity design choice underneath all three
- No child's choice is ever coded as a failure. That is not just a tone commitment — it is what keeps the signal clean. A child who fears being marked "wrong" masks their real strategy; a child who knows every choice is treated as information plays closer to their actual capacity.
Every Interest Area activity (Building, Design, Music, Writing, Animals, Adventure) is separately tagged to a functional domain and an academic skill, with a parallel real-life activity list — so a caregiver or teacher can watch for the same skill outside the app, in a form that isn't asking them to interpret a raw session log.
The Onboarding Assessment (Step 2 of the walkthrough) already names the three domains Hero's Embrace measures directly: Adaptive Function, Executive Function, and Social Function. Metacognition is the process built into every Reflect beat that helps a child actually gain ground in the three domains below: noticing a choice, a feeling, or a strategy honestly enough to build on it next time, without the heavy verbal self-analysis a talk-therapy-style check-in would require. Every Solve-phase choice also asks for some cause-effect judgment, comparison, or sequencing — that reasoning is practiced inside the domains below, not scored as a category of its own.
Adaptive Function
Routines, self-care, transitions, managing belongings, asking for help appropriately.
Where it's practiced
- Keeper's Cove ("The Lighthouse Keeper's Checklist") and Tidewell Flats ("Following the Tide Pictures") — routine and sequencing quests.
- Lanternroot Camp and Windrow Bluff — pacing and self-care over a multi-day in-story stretch, at higher tiers.
What's measured
- The 8-item adult / 5-item child Adaptive Function domain score, weighted most heavily on caregiver report (40%) — the rater with the most daily-living visibility.
- Real-life activity mirrors (chores, self-care routines, packing, safety rules) that caregivers can log against the same domain.
Executive Function
Task initiation, working memory, planning, flexibility, frustration recovery, organization.
Where it's practiced
- Build's Focus & Plan step: sketch a blueprint first, sort a Materials Chest before starting.
- Vinewood Hollow, Emberwatch Rise, and Eastern Ridge quests are anchored specifically to Executive Function.
- "Finding the Row Again" at the Bramble Garden — returning to an interrupted task, deliberately.
What's measured
- The 8-item adult / 5-item child EF domain score at onboarding, re-administered at natural milestones.
- ALO's continuous frustration/disengagement indicator and skill-practice progression trend across the 12-week window (Aim 2's core feasibility metric).
- EF-tagged real-life activities give caregivers a matching thing to watch for at home.
Social Function
Initiating, reading cues, turn-taking, conflict recovery, expressing feelings, awareness of impact on others.
Where it's practiced
- Adventure is the one Interest Area that only works with real peers — welcoming a new camper, noticing an empty seat at the Buttonwood Circle, group negotiation at the Long Table.
- The Companion is designed to never resolve the interaction for the Hero — it stays in earshot so the child works it out themselves.
What's measured
- The 8-item adult / 5-item child Social Function domain score, weighted most heavily on child self-report (35%) of the three domains.
- Caregiver- and educator-reported implementation attempts following SF-tagged Transfer Moment Debriefs — the Aim 3 implementation-feasibility metric.
The SBIR Research Strategy names the core clinical problem directly: children in this population often know what a moment calls for but lack the scaffolding to reliably act on it outside the session where they practiced it. Measuring in-app engagement alone can't answer whether that gap is narrowing — so nothing in this design treats in-app data as the whole picture.
Everything described above — domain scores, Transfer Moment Debriefs, engagement trends — has to reach the right person, in a form scoped to what that person actually needs to see and act on. The Multi-Stakeholder Dashboard prototype models seven such views onto the same underlying data: Caregiver, Teacher, School Social Worker, Counselor, Therapist/Clinician, Medical, and Researcher.
The Researcher persona is built specifically around: a Consent & Governance Status panel (IRB protocol status, active consents, data retention window), a de-identified cohort trends view limited to consenting participants only, an Outcome Pattern Explorer that tracks the CST framework's three falsifiable hypotheses (H1, H2, H3) against real-world outcomes, and equity & representation indicators monitoring access equity across sites.
Three aims, each with a pre-specified (placeholder, pending research-lead sign-off) feasibility threshold — descriptive analysis, not efficacy testing.
| Aim | Question | Feasibility threshold |
|---|---|---|
| Aim 1 Usability & Engagement | Do children engage meaningfully and sustainably over 12 weeks? | ≥60% complete 8/12 sessions · ≥50% voluntarily re-engage · exit interviews predominantly positive/neutral |
| Aim 2 Adaptive Personalization | Does the ALO keep each child in a healthy zone of challenge? | Frustration/disengagement below threshold in ≥80% of sessions · non-negative skill-practice trend for ≥70% of participants |
| Aim 3 Implementation | Can caregivers/educators act on what the platform reports? | Majority of debriefs → a reported home implementation attempt within one week |
BRIEF-2, BASC-3, and Vineland-3 are administered at baseline and study end for descriptive, exploratory purposes only — Phase I is not powered for efficacy testing. Their role here is to begin building construct-validity evidence for the platform's own low-cost functional check-in, so Phase II can test the CST framework's three falsifiable hypotheses (H1 — stress-mediated variability, H2 — Transfer Moment efficacy, H3 — equilibrium stabilization) with instruments already shown to track meaningfully against Hero's Embrace's own domain scores.
| Piece | Status today |
|---|---|
| Placement check-in & domain scoring (Adaptive / Executive / Social) | Specified, sample data live |
| ALO real-time signal processing during play | Specified, not yet built |
| Build → Solve → Reflect Skill Quest content | Conceptual prototype, storyboard only |
| Transfer Moment Debrief | Designed, demo only |
| Multi-Stakeholder Dashboard (Caregiver / School / Clinician / Medical / Researcher) | Structural prototype, no data populated |
| Correlation against BRIEF-2 / BASC-3 / Vineland-3 | Planned for Phase I, not yet run |
This page describes the measurement design the Phase I pilot exists to build and test — not results that already exist. That distinction is the pilot's whole purpose: find out, honestly and rigorously, whether this architecture holds up before asking Phase II to test whether it works.
© 2026 Embracing Neurodiversity. All Rights Reserved. This prototype—including its concepts, design, methodology, and content—is the confidential and proprietary property of Embracing Neurodiversity. It is shared solely for the purpose of evaluation by prospective funders and partners. Unauthorized use, reproduction, distribution, or derivative development is strictly prohibited.