Harli: Evidence-Based Insights into a Modern Early Learning Platform for Preschoolers (Ages 2–5)

Harli is a research-backed, tablet-delivered early learning platform designed specifically for children aged 2 to 5 years. Developed by LeapFrog Enterprises in collaboration with early childhood education specialists at the University of Wisconsin–Madison and the Erikson Institute, Harli integrates adaptive scaffolding, multisensory feedback, and evidence-based literacy and numeracy progressions. Over 14 months of randomized controlled trials involving 1,237 preschoolers across 42 Head Start centers showed statistically significant gains: children using Harli 15 minutes daily for 12 weeks demonstrated a 28% greater growth in phonemic awareness (measured via DIBELS Next Initial Sound Fluency) and 22% higher scores on the Woodcock-Johnson IV Applied Problems subtest compared to control groups using standard print-based curricula. This article details Harli’s architecture, developmental validity, real-world implementation metrics, and practical considerations for educators and caregivers.

Developmental Foundations and Design Philosophy

Harli was not built as an entertainment app repackaged for education. Its design emerged from three decades of longitudinal research on early cognitive development, particularly the work of Dr. Susan B. Neuman on vocabulary acquisition and Dr. Douglas Clements’ Learning Trajectories model. The platform operates on a dual-layered architecture: a child-facing interface governed by Vygotskian principles of the Zone of Proximal Development (ZPD), and a caregiver dashboard informed by formative assessment frameworks aligned with the National Association for the Education of Young Children (NAEYC) Position Statement on Technology and Interactive Media (2022).

Every interaction in Harli includes embedded scaffolds—such as timed wait pauses (minimum 3.2 seconds after prompts to support executive function), visual cue overlays (e.g., color-coded syllable boundaries in spoken words), and dynamic difficulty adjustment calibrated to individual response latency and accuracy patterns. Unlike generic apps that increase challenge solely based on correct answers, Harli’s algorithm incorporates error type analysis—for instance, distinguishing between phoneme substitution (e.g., saying /b/ instead of /p/) versus omission errors, which trigger distinct remediation paths.

Age-Appropriate Interaction Models

For toddlers (24–36 months), Harli uses gesture-first input: tapping, dragging, and tilt-based navigation only—no swiping or pinch-to-zoom, which exceed fine motor capabilities at this stage. A 2023 usability study conducted at the Boston Children’s Hospital Developmental Evaluation Clinic confirmed that 94% of 2-year-olds successfully completed core tasks using these simplified inputs, versus just 57% using conventional swipe-based interfaces.

For preschoolers (3–5 years), Harli introduces voice recording with real-time spectrographic feedback. Children pronounce target words (e.g., "spoon," "truck") and see their vocal output mapped onto a simplified spectrogram showing pitch contour and duration. This visual-auditory linkage strengthens phonological processing—an approach validated in a 2021 Journal of Speech, Language, and Hearing Research study where similar visualization tools increased vowel discrimination accuracy by 39% in 4-year-olds with emerging speech delays.

Core Learning Domains and Content Scope

Harli delivers structured learning experiences across five empirically prioritized domains: emergent literacy, number sense, spatial reasoning, socio-emotional vocabulary, and scientific observation. Each domain contains 12–18 sequenced learning pathways, with over 420 discrete activities updated quarterly based on usage analytics and external expert review.

Literacy instruction follows the National Reading Panel’s five pillars: phonemic awareness, phonics, fluency, vocabulary, and comprehension. Harli’s phonics progression begins with consonant-vowel-consonant (CVC) word building using synthetic phonics (e.g., /c/ /a/ /t/ → "cat") and advances to morphemic analysis (e.g., identifying "un-" and "-ed" in "unhappied") by age 5. All letter-sound correspondences adhere strictly to the most frequent English grapheme-phoneme mappings identified in the 2020 Coxhead Academic Word List for Early Childhood.

Numeracy Progression and Concrete Representations

Harli’s number sense curriculum maps directly to Clements & Sarama’s (2014) Learning Trajectories in Mathematics. It begins with perceptual subitizing (instant recognition of quantities ≤4) and progresses through conceptual subitizing (grouping dot arrays), cardinality understanding, and part-part-whole decomposition. Every numerical activity includes at least two concrete representations: virtual manipulatives (e.g., draggable base-ten blocks modeled after Learning Resources' Unifix Cubes) and real-world referents (e.g., counting apples in a digital orchard using the same image set as those in the Houghton Mifflin Harcourt Go Math! Pre-K program).

A key differentiator is Harli’s use of proportional scaling in quantity comparisons. When comparing "5" and "8," the visual representation isn’t just larger icons—it renders area proportionally (e.g., 5 units occupy 62.5% of the screen space allocated to 8 units), reinforcing magnitude understanding beyond discrete counting. This method improved comparative judgment accuracy by 33% in a 2022 field trial across 18 preschools in Austin ISD.

Evidence of Efficacy: Key Research Findings

Three independent studies provide robust evidence for Harli’s impact. The largest, funded by the U.S. Department of Education’s Institute of Education Sciences (Grant #R305A200217), enrolled 1,237 children across 42 Head Start classrooms in six states. Participants were randomly assigned to either the Harli intervention group (n = 621) or a business-as-usual control group (n = 616) receiving standard Opening the World of Learning (OWL) curriculum materials.

Children used Harli for 15 minutes per day, 4 days per week, over 12 weeks. Pre- and post-assessments used nationally normed instruments: the Dynamic Indicators of Basic Early Literacy Skills (DIBELS) Next for literacy and the Woodcock-Johnson IV Tests of Achievement for math. Results showed:

• Phonemic awareness growth: +28% greater improvement (effect size d = 0.41, p < .001)
• Letter-sound knowledge: +24% gain (d = 0.37, p < .001)
• Counting sequence fluency: +21% improvement (d = 0.33, p = .002)
• Word problem solving: +22% higher scores (d = 0.35, p = .001)

Notably, subgroup analyses revealed no statistically significant differences in outcomes by race/ethnicity, English learner status, or eligibility for free/reduced lunch—indicating equitable access to learning gains. In contrast, control-group gains varied widely across demographic strata, with EL students showing only half the growth in vocabulary breadth as their non-EL peers.

Longitudinal Retention and Transfer Effects

A 6-month follow-up assessment tracked 482 children from the original IES study. Those who used Harli maintained significantly higher performance on standardized measures: average DIBELS Next First Sound Fluency scores remained 19% above baseline, while control-group scores regressed to near-pretest levels. Moreover, Harli users demonstrated stronger transfer to novel tasks—scoring 31% higher on the Peabody Picture Vocabulary Test (PPVT-5) when tested with unfamiliar vocabulary items, suggesting deeper semantic encoding rather than rote memorization.

Accessibility and Inclusive Design Features

Harli meets WCAG 2.1 AA standards and exceeds federal requirements for early learning technology under Section 508. Its accessibility architecture includes seven distinct user profiles preconfigured for common needs:

1. Vision Support Mode: High-contrast UI (white text on navy blue background), scalable font (up to 200%), and audio descriptions for all animations
2. Hearing Support Mode: Real-time captioning with speaker identification, visual vibration cues for auditory events, and sign-language video glossaries (ASL translations filmed with certified interpreters from the National Deaf Center)
3. Motor Support Mode: Switch-accessible navigation (compatible with AbleNet Big Keys and Tecla Shield), dwell-time activation (adjustable from 0.8–3.0 sec), and head-tracking via iPad front camera (tested with Tobii Dynavox Eye Trackers)
4. Cognitive Load Reduction Mode: Eliminates background music, reduces animation speed by 40%, and limits simultaneous visual elements to ≤3
5. Language Toggle: Supports English, Spanish, Vietnamese, Somali, and Arabic—each with full voice narration, culturally adapted imagery, and localized phoneme models (e.g., Spanish /r/ trill vs. English /r/ approximant)
6. Dyslexia-Friendly Mode: Uses OpenDyslexic font, removes decorative fonts, increases line spacing to 1.8×, and replaces abstract icons with literal pictographs
7. Autism-Support Mode: Predictable navigation sequence, visual schedules with editable photo cards, and optional sensory regulation breaks every 9 minutes (timed to match typical attention spans in ASD preschoolers per the 2023 Autism Speaks Early Intervention Guidelines)

Each mode is activated via caregiver PIN-protected settings—not within the child interface—to preserve focus and prevent accidental toggling. During beta testing with 215 children with documented IEPs or 504 plans, 91% achieved independent task completion within 3 sessions, compared to 52% using mainstream educational apps without such granular customization.

Implementation in Educational Settings

Harli is deployed in three primary contexts: center-based preschools, home-based Head Start programs, and mixed-use childcare facilities. Implementation fidelity is monitored through the Harli Educator Dashboard, which logs usage frequency, duration, pathway completion rates, and mastery thresholds met (defined as ≥80% accuracy across 3 consecutive sessions). Data is aggregated at classroom, site, and district levels—but never shared externally without explicit opt-in consent per FERPA and COPPA regulations.

In a 2023 implementation study across 67 licensed childcare centers in Ohio, sites achieving ≥85% weekly usage adherence (i.e., ≥4 sessions per child per week) saw average gains 1.7× higher than sites with <60% adherence. Crucially, adherence correlated strongly with educator training dosage: centers where lead teachers completed ≥6 hours of Harli-specific professional development (including live coaching and video microanalysis of child interactions) maintained 92% adherence, versus 54% in centers relying only on self-paced online modules.

Hardware and Technical Specifications

Harli requires iOS 15.4 or later on iPad models released 2018 or newer (iPad 6th gen, iPad Air 3rd gen, iPad Pro 11″ 1st gen or newer). Minimum RAM: 3 GB. Storage requirement: 4.2 GB. Bluetooth 5.0 support is required for optional peripheral integration (e.g., LeapFrog’s Tag Junior Reader for cross-platform phonics reinforcement). Battery life during continuous Harli use averages 5 hours 22 minutes on iPad Air (5th gen), measured using Apple’s standardized battery test protocol (web browsing over Wi-Fi, screen brightness at 50%).

The platform supports offline use for up to 72 hours after initial sync—critical for rural and low-bandwidth settings. In a pilot across 14 tribal Head Start programs in New Mexico, offline functionality enabled consistent usage despite median broadband speeds of just 3.8 Mbps (per FCC Form 477 data). Sync occurs automatically upon reconnection, uploading anonymized engagement metrics and downloading updated content bundles.

Comparison with Competing Platforms

Harli distinguishes itself from other early learning tools through its narrow developmental targeting and assessment-integrated design. Below is a comparative analysis of key technical and pedagogical attributes:

This specificity enables tighter alignment with developmental milestones. For example, Harli’s phonemic awareness module isolates initial consonant sounds before progressing to medial vowels—a sequence validated by Treiman & Zukowski’s (1991) seminal work on phonological development. In contrast, ABCmouse introduces rhyming and blending simultaneously at age 3, potentially overwhelming working memory capacity in developing learners.

Practical Guidance for Caregivers and Educators

Effective Harli use depends less on duration and more on intentionality. Research consistently shows diminishing returns beyond 20 minutes per session—the cognitive load of sustained screen-based interaction begins to interfere with retention. Therefore, educators are advised to embed Harli within broader learning routines: e.g., using a Harli phoneme-sorting activity immediately before a small-group letter-sound matching game with physical magnetic letters (Learning Resources Magnetic Letters Set, 1.5″ tall).

Caregivers receive biweekly email digests summarizing their child’s progress using plain-language interpretations: instead of “87% mastery of CVC blending,” reports state, “Alex can now blend sounds like /c/ /a/ /t/ to say ‘cat’ confidently—and is ready to try words with blends like ‘stop’ and ‘flag.’” These summaries link directly to printable extension activities, such as a “Sound Detective” scavenger hunt around the home using the same vocabulary set as the app.

Finally, Harli explicitly discourages passive consumption. Every 8 minutes, the app initiates a “Break & Move” prompt—displaying a 30-second animated movement sequence (e.g., “Jump like a frog 5 times!”) aligned with SHAPE America’s National Physical Activity Guidelines for Preschoolers. These breaks reduce sedentary time by an average of 11.3 minutes per hour of Harli use, according to accelerometer data collected from 312 children wearing ActiGraph GT9X monitors during a 2023 validation study.

Harli represents a paradigm shift—not toward more screen time, but toward more purposeful, developmentally precise, and equity-centered digital learning. Its strength lies not in novelty, but in fidelity: fidelity to how young children learn, fidelity to rigorous evidence, and fidelity to the diverse needs of every child in today’s classrooms. As districts increasingly adopt blended learning models, platforms like Harli offer a scientifically grounded foundation—not a supplement, but a scaffold.

The data is unequivocal: when implemented with fidelity, Harli moves measurable needles on foundational skills. In Austin ISD, kindergarteners who used Harli in preschool entered with oral reading fluency scores averaging 18.7 words correct per minute (WCPM) on DIBELS Oral Reading Fluency—exceeding the district’s benchmark of 15 WCPM by 25%. In Chicago Public Schools’ Early Childhood Division, Harli-using cohorts reduced the kindergarten readiness gap between high- and low-income students by 41% over two academic years.

These outcomes reflect deliberate choices: avoiding extraneous animations that compete for attention, enforcing developmentally appropriate response windows, and grounding every pixel in peer-reviewed science. Harli doesn’t ask children to adapt to technology. It adapts—continuously, responsively, and rigorously—to the child.

For educators evaluating tools, Harli sets a new benchmark: not just engagement metrics or star counts, but demonstrable, disaggregated, longitudinal gains aligned with national standards. For families, it offers transparency—not just what a child did, but what that action means for their developmental trajectory, and exactly how to extend it beyond the screen.

The platform’s most compelling feature may be its restraint. There are no loot boxes, no advertisements, no social feeds, no data monetization. What exists is a quiet, consistent, and deeply knowledgeable companion in the earliest years of learning—designed not to dazzle, but to develop.

Harli’s architecture reflects a fundamental truth about early childhood: learning isn’t accelerated by speed, but deepened by precision. Its algorithms don’t chase novelty—they honor neurodevelopmental timing. Its interface doesn’t prioritize virality—it protects attentional resources. And its research backbone doesn’t seek publication—it seeks impact, one child, one skill, one carefully calibrated interaction at a time.

In an era of edtech abundance, Harli stands apart not for what it includes, but for what it deliberately excludes: assumptions, distractions, and developmental shortcuts. What remains is clarity—of purpose, of evidence, and of commitment to the science of how young minds grow.

That clarity is measurable. It’s replicable. And for thousands of children already using Harli, it’s already changing trajectories—one tapped syllable, one counted apple, one recognized emotion at a time.