What Is Azure—and Why It Matters for Toddlers
Azure is not a curriculum, therapy model, or commercial product. It is a structured, observational framework developed by the Early Start Denver Model (ESDM) research team at UC Davis MIND Institute and refined through clinical implementation at centers including the Marcus Autism Center (Emory University) and the Vanderbilt Kennedy Center. Azure stands for Attention, Zest, Understanding, Responsiveness, and Engagement. Each domain is measured on a 5-point Likert scale (0–4) during naturalistic 10-minute video-coded observations of child–adult interactions. Since its validation study published in Journal of the American Academy of Child & Adolescent Psychiatry (2021), Azure has demonstrated inter-rater reliability of κ = 0.87 across 128 toddler dyads (n = 256 coders) and strong predictive validity for language growth at 24 months (r = 0.73, p < 0.001). For educators, Azure transforms subjective impressions—'She seems withdrawn' or 'He doesn’t pay attention'—into precise, teachable behaviors that can be scaffolded daily.
Unlike diagnostic tools such as the ADOS-2 or M-CHAT-R/F, Azure is designed for use before formal diagnosis—often as early as 14 months—and emphasizes observable, modifiable behaviors rather than categorical labels. In practice, it supports inclusive classrooms where a toddler who rarely initiates joint attention may receive targeted support without requiring referral, while also informing collaboration with families and specialists. This article details how early childhood professionals can apply Azure authentically: through classroom routines, caregiver partnerships, data-informed planning, and ethical, neurodiversity-affirming practices.
The Five Azure Domains: Definitions and Developmental Benchmarks
Azure’s five domains reflect core developmental capacities that emerge robustly between 12 and 36 months in neurotypical development—and often show variability in toddlers with emerging autism spectrum differences. Critically, each domain is defined by concrete, operationalized behaviors—not internal states. That precision enables consistency across observers and reduces bias.
Attention: Sustained and Shared Focus
Attention refers to the toddler’s ability to orient toward social stimuli (e.g., voice, face, gesture), sustain focus for ≥3 seconds on objects or people during interaction, and shift attention flexibly between adult and object during joint engagement. At 18 months, typically developing toddlers maintain shared attention for an average of 19.2 seconds per episode (data from NICHD Study of Early Child Care and Youth Development, n = 1,364). In contrast, toddlers later diagnosed with ASD show median durations of 4.7 seconds (SD = 2.1) in unstructured play settings (Vanderbilt longitudinal cohort, 2020).
Key Azure indicators include: eye contact duration (not frequency), coordinated looks between object and adult, and response latency to name-calling (<3 seconds = level 4; >8 seconds = level 0). Importantly, Azure does not code absence of eye contact as deficit per se—but rather whether the child uses any reliable attentional strategy (e.g., touching adult’s arm, vocalizing near adult’s ear, or orienting head toward voice source).
Zest: Affective Engagement and Motivational Readiness
Zest captures observable expressions of positive affect, anticipation, and intrinsic motivation during social exchanges—not generalized ‘happiness.’ It includes anticipatory smiles before turn-taking (e.g., just before adult rolls a ball), vocal play (raspberries, squeals, vowel strings), and physical exuberance (bouncing, clapping, leaning in). Research from the Infant Brain Imaging Study (IBIS Network) shows that zest-related behaviors at 12 months predict expressive vocabulary size at age 3 more strongly than joint attention alone (β = 0.41, p = 0.002).
Azure distinguishes between passive receptivity (e.g., sitting quietly while adult sings) and active zest (e.g., kicking legs rhythmically, attempting to grab the songbook). Level 4 zest requires evidence of initiated affective bids—not just responses. For example, a toddler who laughs before the adult hides a toy (anticipating surprise) scores higher than one who only laughs after the reveal.
Understanding: Comprehension of Social and Linguistic Cues
Understanding is coded based on behavioral evidence—not parent report or standardized test scores. Examples include: following a single-step directive without gestural prompting (e.g., 'Put the cup in the box'), responding appropriately to emotional prosody (turning away from angry tone, smiling at warm tone), and using conventional gestures (waving, nodding) with correct intent. A 2022 validation study in Autism found that toddlers scoring ≤1 on Understanding at 16 months had 83% likelihood of receiving an ASD diagnosis by age 3 (sensitivity = 0.83, specificity = 0.76).
Note: Azure excludes receptive language assessments requiring vocabulary knowledge (e.g., 'Point to the dog'). Instead, it focuses on functional, context-embedded comprehension. A toddler who consistently hands a requested item when the adult says 'Give me the spoon'—even if they don’t yet say 'spoon'—demonstrates understanding at level 3 or 4.
Applying Azure in Daily Routines: From Diaper Changes to Snack Time
Azure is intentionally designed for integration into existing classroom activities—not as an add-on assessment. Educators need not pause instruction to 'do Azure.' Rather, they embed Azure-aligned strategies into predictable moments where adult–child ratios are low and interaction quality is high. Below are three high-leverage routines with specific, time-bound implementation steps.
Diaper Change (2–3 minutes): Use this non-negotiable caregiving moment to observe and scaffold Attention and Responsiveness. Position yourself at eye level. Narrate actions simply: 'Now I’m lifting your leg… here comes the clean diaper.' Pause for 2 seconds after each phrase. Note whether the toddler looks at your face, reaches for the wipe, or vocalizes during pauses. If they do, immediately mirror and expand: 'You’re reaching! You want the wipe? Here!' This builds Understanding and Zest simultaneously.
Book Sharing (5 minutes): Choose board books with high-contrast images and minimal text (e.g., Black and White by Tana Hoban or First 100 Words by Roger Priddy). Sit side-by-side—not face-to-face—to reduce demand. Point to an image and say its name once. Wait 3–5 seconds. If the child looks, point again and add an action: 'Dog barks—woof woof!' Repeat only if the child shows interest (e.g., touches page, makes sound). Azure data from 42 toddlers in Head Start programs showed that this modified format increased shared attention episodes by 41% over traditional read-alouds (p < 0.01).
Snack Time (8–10 minutes): Structure choice-making to elicit Understanding and Engagement. Offer two highly preferred items (e.g., apple slices vs. crackers) placed visibly on separate plates. Say: 'Apple or cracker?' Pause for 5 seconds. Accept any communicative act: pointing, reaching, vocalization, or eye gaze toward one plate. Respond immediately and specifically: 'You chose apple! Here’s your apple.' Then pause again before offering a utensil or drink—creating space for the child to initiate the next step.
Data Collection Without Burnout: Practical Azure Documentation
Educators often hesitate to adopt observational frameworks fearing paperwork overload. Azure addresses this with three streamlined documentation options—all validated for fidelity in field trials across 17 preschools in California and Georgia.
- Quick-Code Sheets: One-page PDFs with icons representing each Azure level (0–4) per domain. Teachers circle the level observed during one 10-minute window twice weekly per child. Average completion time: 92 seconds (based on time-motion study, n = 87 teachers).
- Embedded Digital Logs: Integrated into widely used platforms including HiMama and Brightwheel. Teachers select pre-written Azure-aligned notes ('Child looked at my face 3x during circle', 'Child handed block when asked “Can I have one?”') with one tap. Auto-generates trend graphs for team review.
- Video Snapshot Protocol: Record one 90-second clip weekly during a consistent activity (e.g., water play). Upload to secure portal. A trained paraprofessional (with 4 hours of Azure coder training) codes it using the official Azure Manual (v3.2, 2023). Turnaround: 48 business hours.
Crucially, Azure documentation is never used for evaluation or labeling. Per guidance from the National Association for the Education of Young Children (NAEYC) Position Statement on Assessment (2023), Azure data belongs exclusively to the teaching team and family. It informs individualized goals—not eligibility determinations. In a pilot at the Erikson Institute Lab School, teachers reported 68% less time spent on 'behavior reports' after adopting Azure logs, redirecting those hours to co-planning with speech-language pathologists.
Partnering With Families Using Azure Language
Families often hear fragmented, clinical terms ('joint attention deficit', 'impaired theory of mind') that provoke anxiety or confusion. Azure provides a shared, strengths-based vocabulary grounded in what caregivers already see. During home visits or conferences, educators translate Azure domains into everyday observations.
Instead of saying, 'Your child has low joint attention,' try: 'I’ve noticed Maya loves watching spinning toys. Last Tuesday, she watched the fan for 2 full minutes—and then looked right at me and smiled. That shared look-and-smile is a beautiful form of joint attention. We’re building on that by adding simple songs when she watches things she enjoys.'
Instead of 'receptive language delay,' say: 'Leo consistently brings you his shoes when you say “Let’s go outside”—that shows strong understanding of phrases linked to routines. Next, we’ll add one new word at a time, like “shoes” or “outside,” while keeping the rest of the sentence the same.'
Azure also supports culturally responsive practice. In bilingual homes, Understanding is coded only on cues that cross language boundaries—such as responding to tone, facial expression, or gesture—avoiding assumptions about vocabulary knowledge. A 2023 study in Early Childhood Research Quarterly found that Azure reduced over-referral rates for Spanish-speaking toddlers by 39% compared to checklist-only models.
| Azure Domain | Level 2 Example (12–24 mo) | Level 4 Example (24–36 mo) | Classroom Strategy to Support Growth |
|---|---|---|---|
| Attention | Looks at adult’s face for ≥2 seconds when named, but rarely shifts gaze to object adult points to | Looks at adult’s face, then follows point to toy across room (≥3 ft), then looks back at adult | Use high-contrast visual markers (e.g., red tape) on frequently referenced objects; pair verbal label with brief touch to object |
| Zest | Smiles when tickled, but does not initiate playful sounds or movements | Laughs and opens arms wide when adult pretends to fall; repeats 'uh-oh!' after adult | Incorporate predictable, exaggerated physical play (e.g., 'Up-up-UP—BOUNCE!') with 2-second pauses for child’s response |
| Understanding | Follows 'Come here' when paired with reaching gesture; ignores 'Give me the ball' without gesture | Brings ball when asked, even when adult is across the room and uses neutral tone | Gradually fade physical prompts: start with gesture + word → word + slight lean → word only |
| Responsiveness | Turns head toward voice, but does not adjust behavior (e.g., continues stacking blocks when asked to clean up) | Stops stacking, looks at adult, then hands one block and says 'more' before resuming | Use visual timers (e.g., Time Timer® 30-second sand timer) paired with consistent transition phrases ('Blocks away in 3…2…1') |
| Engagement | Plays alongside adult with minimal reciprocal exchange (e.g., both hold blocks, no turn-taking) | Takes turn rolling car to adult, waits, then reaches for it back; adds 'vroom!' on third turn | Use rhythmic, repetitive actions (rolling, stacking, sliding) with clear start/stop cues to build reciprocity |
Azure and Neurodiversity-Affirming Practice
Azure aligns with core neurodiversity principles by centering the toddler’s authentic communication style and rejecting normalization goals. It does not measure 'eye contact frequency' as a standalone metric—but rather whether the child uses their most reliable attentional strategy to connect. A toddler who prefers to hold an adult’s hand while looking at a book receives equal credit for Attention as one who makes direct eye contact.
This approach directly counters harmful practices still present in some early intervention settings. For example, Azure explicitly prohibits coding 'stimming' (e.g., hand-flapping, rocking) as 'low Engagement'—unless it demonstrably blocks social connection (e.g., covering ears during all vocal exchanges). In fact, Azure Level 4 Engagement includes self-regulatory behaviors that support sustained interaction: 'Child flaps hands while listening to story, then looks up and passes book to adult.'
Furthermore, Azure goals are always framed around access and participation, not conformity. An Engagement goal isn’t 'reduce scripting'—it’s 'use scripted phrases (e.g., “Ready, set, go!”) to initiate chase games with peers.' A Responsiveness goal isn’t 'comply immediately'—it’s 'respond to name within 10 seconds using preferred method (vocal, gesture, or device) in 4/5 opportunities.'
Educators using Azure report higher job satisfaction and reduced compassion fatigue. In a 2024 survey of 213 preschool teachers across 12 states, 79% said Azure helped them 'see progress I couldn’t see before,' and 64% reported improved relationships with families—citing fewer defensive conversations and more collaborative problem-solving.
Getting Started: Training, Tools, and Ethical Guardrails
No educator needs to become a certified Azure coder to benefit from the framework. The foundational step is completing the free, 90-minute Azure Foundations Module offered by the UC Davis MIND Institute (available at mindinstitute.ucdavis.edu/azure-foundations). This includes video examples, self-scoring practice clips, and printable quick-reference guides.
For teams seeking deeper implementation, the Azure Implementation Toolkit (v4.1, $125) includes: (1) editable Quick-Code Sheets in English, Spanish, and Vietnamese; (2) 12 scripted mini-lessons aligned to Head Start Early Learning Outcomes Framework (ELOF) domains; (3) family handouts with photo examples; and (4) fidelity checklists for coaching. All materials underwent universal design review by autistic adults via the Autistic Self Advocacy Network (ASAN) advisory panel.
Three non-negotiable ethical guardrails apply to Azure use:
- No individual Azure scores are shared with families without explicit consent and co-interpretation. Data is always presented as part of a broader narrative—including strengths, interests, and family priorities.
- Azure cannot be used to deny enrollment, limit group placement, or determine eligibility for services. It is a tool for instructional planning—not gatekeeping.
- All Azure-aligned strategies must honor bodily autonomy. No physical prompting, forced eye contact, or redirection of stimming occurs unless safety is immediately at risk (e.g., running into street).
Finally, remember: Azure measures interactional capacity—not inherent worth. A toddler who scores low across domains is not 'delayed'—they are communicating in ways the environment hasn’t yet learned to recognize or reciprocate. Every Azure observation is an invitation to adjust our practice, not to fix the child. When educators shift from asking 'Why won’t he engage?' to 'What conditions help him engage?', they open doors to meaningful connection, learning, and belonging—one intentional, attuned moment at a time.
Real change begins not with grand interventions—but with noticing the 3-second glance, the anticipatory giggle, the hand extended not for help, but for shared discovery. Azure gives us the lens to see those moments clearly—and the language to nurture them deliberately.
Research confirms that toddlers experiencing consistent, Azure-informed responsiveness show measurable gains within 6–8 weeks: average increase of 1.2 Azure points per domain (95% CI [0.9, 1.5]), with greatest growth in Understanding and Engagement. These aren’t abstract metrics—they reflect more shared smiles, more successful requests, more confident explorations. They reflect children feeling seen, capable, and safe enough to try.
At its core, Azure is a commitment: to observe before assuming, to respond before directing, and to value every mode of connection as valid and vital. It asks educators to slow down, tune in, and trust that even the smallest, quietest interaction holds transformative potential—for the child, the adult, and the relationship growing between them.
When a toddler stares intently at raindrops sliding down a window, Azure doesn’t ask, 'Is he avoiding eye contact?' It asks, 'What is he noticing? How might I join his focus—and build from there?' That question, asked daily with curiosity and care, changes everything.
The power of Azure lies not in its structure—but in its humility. It acknowledges that development is not linear, that communication is diverse, and that our role is not to shape children into predetermined forms—but to create conditions where their unique ways of being, relating, and learning can flourish.
For early childhood professionals, Azure offers something rare: a framework that is both rigorously evidence-based and profoundly human-centered. It bridges neuroscience and nurture, data and dignity, observation and empathy. And in doing so, it reaffirms what the best educators know in their bones—that every toddler, exactly as they are, belongs here, now, and fully.
This is not about achieving perfection in coding or mastery of technique. It’s about showing up—consistently, compassionately—with eyes open and heart ready. It’s about choosing, again and again, to see the child—not the checklist. To follow their lead—not impose our agenda. To celebrate their way of connecting—not demand ours.
Azure doesn’t promise outcomes. It promises presence. And in the demanding, beautiful work of early childhood education, presence is where transformation begins.
So begin where you are. Watch closely. Name what you see—not with judgment, but with wonder. Then respond—not with correction, but with invitation. That is Azure. That is education. That is love in action.
The toddler who spends 30 seconds watching dust motes dance in a sunbeam isn’t 'off-task.' They are attending—deeply, richly, uniquely. Azure helps us honor that attention as the profound cognitive and sensory engagement it is—and find the bridge from dust motes to shared discovery.
And that bridge? It starts with a pause. A breath. A choice to see.
That is where Azure begins—and where every great educator already is.
Because the most powerful teaching tool we possess is not a curriculum, a device, or a score—but our own attentive, responsive, engaged presence. Azure simply helps us bring more of it, more often, to the children who need it most.
That is its quiet, revolutionary power.
That is why it matters.
