The EdTech Comparison: What Schools Have, What Failed, and What's Different Now

The Comparison That Should Have Been Done Twenty Years Ago: EdTech Approaches, What Failed, and What's Actually Different Now

A side-by-side examination of five educational technology approaches — measured against the research-validated principles that determine whether any tool actually produces learning.

Series: 1 2 3

Doug Hamilton

Pastor, Board Certified Christian Counselor • Founder, Derech Technologies LLC

Christian Pastor Board Certified Christian Counselor AI Developer

Standard Disclosures

Doug Hamilton is a Christian pastor and Board Certified Christian Counselor. His faith informs his worldview. This lens is acknowledged, not hidden.

Organizational Lens Principle — Required Disclosure: Doug is the developer of Spirit-Bridge, one of the five platforms reviewed in this analysis. That financial and vocational stake is declared openly so readers can account for it and weigh the analysis accordingly.

Pre-Launch Disclosure: Spirit-Bridge is evaluated here as a complete designed system — reflecting the platform as it will exist at launch, not its current development state. All features attributed to Spirit-Bridge in this matrix are either fully built, in final development, or committed for completion prior to public release. The four competing platforms are evaluated in their current live operational state. This distinction is important and readers should account for it. We are comparing a finished design against finished products — and we are saying so plainly.

This analysis was produced collaboratively with AI research tools. The methodology, judgment, and conclusions are Doug's. The research breadth is AI-assisted.

No matter how diligently we work to set aside bias, a lens remains. Do your own research. Test these findings. Hold us to our own standard. Proverbs 18:17 applies to us too.

The Context

Before We Compare Anything — We Need a Standard

The current conversation about educational technology keeps comparing tools to each other. That's the wrong comparison. The right comparison is: which approach actually aligns with what the research says about how human beings learn?

The peer-reviewed evidence has converged on a specific set of principles. These aren't new ideas — Vygotsky's Zone of Proximal Development, Bloom's taxonomy, the Ebbinghaus forgetting curve, and John Hattie's synthesis of 800+ meta-analyses — each covering dozens of studies, totaling over 50,000 individual studies — are decades old. The question that every generation of technology has failed to answer is whether the tool can deliver those principles at scale. The headlines calling technology a failure aren't wrong that something failed. They're wrong about what failed.

"The one who states his case first seems right, until the other comes and examines him."

Proverbs 18:17 — The Anchor of This Work

The pandemic pressure-tested a system that was already fundamentally flawed — and what broke wasn't the technology. It was the assumption that had been wrong since the beginning: that access equals outcome.

The Five Approaches

What We're Comparing — and Why These Five

We selected five distinct approaches representing the full arc of educational technology: what schools have done, what currently exists, and what's now emerging. Each is evaluated against the same evidentiary standard — not against each other.

Approach 1

Traditional Public School + Standard EdTech

Chromebooks, Google Classroom, LMS platforms — the current baseline

What virtually every American school uses today: devices distributed, content digitized, curriculum largely unchanged. Provides irreplaceable human teachers, socialization, extracurriculars, special education services, and legal protections. The technology layer, however, has not improved core learning outcomes — and that specific failure is what this comparison examines.

Approach 2

Adaptive Practice Platforms

IXL Learning, DreamBox — the current mainstream "smart" tools

Adaptive difficulty, standards-aligned practice, real-time analytics. A genuine improvement over passive screen time — but fundamentally supplementary, not transformational.

Approach 3

Khan Academy / Khanmigo

AI-powered tutoring — the mainstream emerging model

GPT-4 powered Socratic tutor with 700,000 student users. Genuinely different in approach — asks rather than tells. Early results promising; independent RCT not yet completed.

Approach 4

Alpha School

2-Hour Learning model — the cutting-edge private experiment

Core academics in 2 hours daily via adaptive learning software (similar in technology to IXL, not large language models); afternoons for life skills. Top MAP test results reported. Significant caveats: $40-75K/year tuition, no independent verification of outcomes.

Approach 5

Spirit-Bridge

Full AI learning ecosystem — the emerging principled model

Purpose-built AI tutoring platform with biblical integration, companion-based learning, character development tracking, and a full-system architecture built around learning science from the ground up.

The Pedagogical Comparison Matrix

Twelve criteria drawn from peer-reviewed learning science. Applied consistently to all five approaches. Not graded on a curve.

Fully Present

Partially / Emerging

Absent

	Traditional + Standard EdTech	IXL / DreamBox	Khan Academy / Khanmigo	Alpha School	Spirit-Bridge
Learning Science Core
Mastery-Based Progression Students advance on demonstrated understanding, not calendar time	✗ Time-based. Students advance regardless of mastery.	~ Adjusts difficulty but does not gate curriculum advancement	~ Requires explanation, not just answer selection — emerging mastery model	✓ 90% mastery required to advance. Bloom's 2 Sigma applied.	✓ 80% to advance, 90% = mastered, 95% = exemplary. No time-based advancement. Enforced in code.
Adaptive Personalization Real-time adjustment to individual pace, style, and gaps	✗ One-size-fits-all instruction. Teaching to the median.	✓ Strong. Every interaction informs difficulty and path.	~ Adapts tone and explanation style; less adaptive on curriculum path	✓ Knowledge Graph + Interest Graph per student. Lessons personalized to interests. Note: Alpha's technology uses adaptive learning apps (similar to IXL), not large language models — a meaningful distinction from LLM-based tutors.	✓ VARK learning styles, Vygotsky's ZPD (70–95% challenge zone), interest-matched content, real-time adjustment. Hardcoded into AdaptiveLearningEngine.
Spaced Repetition / Retention Science Review at optimal intervals based on Ebbinghaus forgetting curve	✗ Content reviewed by unit schedule, not retention need	~ DreamBox uses spaced review; IXL loops struggling topics	✗ Not a documented feature of Khanmigo's current methodology	~ Mastery pacing implies review cycles; not explicitly published	✓ Ebbinghaus intervals (1, 3, 7, 14, 30, 90 days) explicitly implemented. Interleaving: 70% current / 30% review. Enforced in MasteryTracker and AdaptiveLearningEngine.
Socratic / Guided Discovery Teaching through questions, not answer delivery — student must explain in own words	✗ Lecture delivery. Teacher presents; students receive.	✗ Practice and feedback, not guided inquiry. Still answer-selection based.	✓ Core design philosophy. Asks questions; cannot give direct answers. Modeled on Socratic dialogue explicitly.	~ Guides toward correct answers; Socratic dialogue not the explicit foundation	✓ Discover → Retrieve → Apply pipeline. Finding answer is Step 1, NOT mastery. Student must explain in own words to advance. Verbatim copy-paste detection. Enforced in prompt engine.
Student Experience
Emotional & Wellness Responsiveness Detects frustration, anxiety, disengagement — responds before content continues	✗ Teacher may detect; no systematic response. One teacher, 28 students.	~ Frustration detection triggers difficulty reduction; no emotional responsiveness	~ Designed to be encouraging; no published emotional state detection system	~ Guides are present and responsive; AI component not emotionally adaptive	✓ Teaching Principle #1 (overrides everything): lesson STOPS on emotional distress. Monitors for negative self-talk, short answers, resistance. Responds with genuine warmth before resuming. Enforced in AI prompt at highest priority.
Student Autonomy & Pacing Control Student can explore, skip ahead, set direction — not locked into linear sequence	✗ Curriculum sequence is fixed. All students move together.	~ Student moves at own pace within skill sequence; path is still predetermined	~ Student can redirect conversation; curriculum sequence is still teacher-assigned	✓ Self-paced by design. Students move through content at their own rate.	✓ Teaching Principle #7: if student wants to explore a tangent, LET THEM. If they already know content, let them prove it and skip. "We need to move on" is explicitly prohibited in the AI instructions.
Character / Whole-Person Development Tracks and develops virtue, character, and spiritual growth — not just academic metrics	✗ Academic metrics only. Character is a separate and often neglected domain.	✗ Academic practice only. No character dimension.	✗ Academic tutoring only. No character tracking.	✗ Life skills workshops in afternoons but no structured character development tracking	✓ CharacterGrowthTracker: Fruit of the Spirit (9 traits) + Universal Virtues (9 traits). 13-year longitudinal tracking woven into every lesson. GiftDiscoveryEngine identifies spiritual, relational, intellectual, artistic, and practical gifts across 20+ data points. No other platform in this comparison attempts whole-person development at this depth.
System, Safety & Transparency
Parent & Teacher Visibility / Reporting Detailed, actionable reporting on what student actually learned and how — for both parents and classroom teachers	~ Gradebook + parent-teacher conferences. Delayed, often summary-level only.	✓ Strong analytics dashboard. Real-time insights, skill gap identification, small group flags.	~ Teacher lesson planning tools are strong; session-level student reporting not yet robust	~ MAP test results reported; session-level transparency limited	✓ 10-section parent report per session: mastery breakdown, engagement profile, emotional moments, conversation highlights, personalized recommendations, next lesson preview. Dedicated teacher dashboard. Intervention flags delivered with specific research-backed strategies — not generic alerts.
Safe, Distraction-Free Environment Student cannot access social media, YouTube, or off-task content during learning	✗ Internet access is the problem. Same device = same temptation.	~ Internet required; filtering is school's responsibility	~ Internet required; anti-cheat measures built in but not distraction control	~ School environment controls; student devices still connected to internet	✓ Local network deployment. Students connect to Spirit-Bridge server only. Cannot access YouTube, social media, or internet. Structurally impossible — not policy-dependent.
Full Curriculum (Not Supplementary) Can function as primary instruction across core subjects and grades	✓ Full curriculum — but built on the assumption that access equals outcome. Note: Google Classroom, the most widely used LMS, functions as a digital filing cabinet for assignments — it organizes content delivery but does not teach.	✗ Explicitly supplementary. Designed to support a teacher, not replace poor instruction.	~ Full subject coverage but designed as tutoring support, not primary instruction	✓ Full K-12 curriculum in core subjects delivered primarily through adaptive software. Note: "AI" here refers to adaptive learning apps, not large language models.	✓ Original curriculum: Math, Reading, Writing, Science, History, Bible, Art, Music. Grades 5-7 + Life Readiness. Hybrid mode: imports publisher curriculum AND generates AI supplements. Three operating modes: Import, Generate, Hybrid.
Access & Affordability
Price & Affordability Cost to families and schools relative to what is delivered — and who can realistically access it	✓ Free to families (taxpayer funded). Device cost absorbed by school. Most accessible by default — but delivers the weakest pedagogical result per dollar spent.	✓ IXL: $5–$12/student/year for schools. DreamBox: ~$20/student/year. Very affordable supplementary tools — strong value for what they are.	✓ Khanmigo: $4/month or $44/year for students. Teacher tools free. The most affordable AI tutoring platform available by a wide margin.	✗ $40,000–$75,000/year tuition. Not a scalable equity solution. MIT researcher noted results cannot be fairly compared to general student population given the resource gap.	~ Homeschool co-op: ~$1,200/year per family — vs. $5,200–$15,600/year for a private tutor. School deployment: server ($4,100 one-time) + per-student licensing. More expensive than IXL or Khanmigo; dramatically less than Alpha or private tutoring. Value case is strong — affordability is moderate.
Teacher Integration Model AI handles content delivery and practice; teacher is liberated for mentorship, critical thinking, and relationship — the things only humans can do	✗ Technology is layered onto the existing teacher role, not redesigned around it. Teacher workload increases, not decreases.	~ Helps teachers identify gaps and differentiate. Does not fundamentally change the teacher's instructional role.	~ Khanmigo for Teachers offers lesson planning and assessment tools — useful, but the teacher's instructional delivery role is not fundamentally restructured. Strong teacher support; not a teacher liberation model.	~ "Guides" oversee students rather than teach. Model moves away from teacher, not toward a redefined hybrid role.	✓ AI handles content delivery, retrieval practice, mastery tracking, and emotional monitoring. Teacher receives session reports, intervention flags with research-backed strategies, and is freed for mentorship and relationship. Designed as a hybrid model, not a teacher replacement.
Overall Score (out of 12 criteria)	2 / 12	5 / 12	6 / 12	7 / 12	11.5 / 12 *
* Spirit-Bridge score reflects the complete designed system at launch. See Pre-Launch Disclosure above. Outcome data is directionally strong and emerging — peer-reviewed longitudinal studies are not yet published.

Strengths, Weaknesses & Evidentiary Tier

Every approach — including Spirit-Bridge — evaluated against our own standard. Proverbs 18:17 applies to us too.

Approach 1

Traditional Public School + Standard EdTech

Chromebooks, Google Classroom, LMS, 1:1 devices

What It Gets Right

Human teacher relationships remain irreplaceable for many students
Physical and social environment supports development that screens cannot
Infrastructure exists — every school already has it
Teachers with high implementation support DO produce meaningful gains

Structural Failures

Operates on the assumption access equals outcome — the assumption that has failed since 1937
Teaching to the median: advanced students bored, struggling students lost, simultaneously
87% of principals say tech is crucial; only 18% say teachers are proficient at using it
Same device for learning AND distraction — structurally compromised
Budget and bureaucracy ensure the gap between technology's capability and classroom use keeps widening

Evidence: Tier 2 — Interpretation Required The device deployment data is Tier 1 verified. Whether the solution is reform or replacement is genuinely contested.

Approach 2

Adaptive Practice Platforms

IXL Learning • DreamBox by Discovery Education

What It Gets Right

DreamBox: ESSA-rated "Strong" — 20% math proficiency boost, 1.6 years growth in one school year for consistent users
IXL: 17,000+ skills, aligned to all 50 state standards, real-time insight for teachers
Genuine adaptive difficulty — not just harder/easier but path-adjustment
Democratizes quality practice for under-resourced schools

What It Misses

Explicitly supplementary — practice platform, not instruction. Cannot replace a teacher or a curriculum.
No Socratic dialogue — still answer-selection, not guided discovery
No emotional responsiveness — struggling students get harder problems, not a warm check-in
No character development, no parent reporting on how the child learns — only what they got right
Gamification can backfire: chasing points, not understanding

Evidence: Tier 1 — Verified (DreamBox) / Tier 2 (IXL) DreamBox has peer-reviewed and ESSA research. IXL's adaptive claims are well-documented but effect sizes vary by implementation.

Approach 3

Khan Academy / Khanmigo

AI Socratic Tutor • 700,000 K-12 users (2024-25)

What It Gets Right

Genuinely Socratic by design: asks, doesn't tell. Modeled explicitly on Socratic dialogue.
Cannot write essays for students — built-in academic integrity
Pre-AI Khan: 30 min/week correlated with greater-than-expected standardized test gains
$4/month — accessibility is real and meaningful
Massive rapid adoption signals something is working

What It Misses

No independent RCT on Khanmigo specifically — honest gap they acknowledge
Student engagement problem: many students type "I don't know" repeatedly rather than engaging
No emotional state detection — doesn't stop for a struggling student's distress
No character development, no companion relationship, no whole-person tracking
Requires internet access — same distraction risks as other browser-based tools

Evidence: Tier 3 — Promising, Pre-RCT Early indicators are strong. The independent randomized controlled trial is planned but not yet published. This is intellectually honest to name.

Approach 4

Alpha School

2-Hour Learning • $40,000–$75,000/year tuition

What It Gets Right

Bloom's 2 Sigma Problem applied: 90% mastery required before advancement
Knowledge Graph + Interest Graph per student — personalization is genuine
Reports top 1-2% MAP test performance in 2024-25
Proves the 2-hour academic day model is viable — afternoons reclaimed for life skills
Interdisciplinary units reflect how real learning actually connects across subjects

What Requires Honest Caveats

Technology clarification: Alpha's "AI" is adaptive learning software similar to IXL — not large language models. The term "AI school" in media coverage overstates the technology's nature.
No independent peer-reviewed verification of outcomes — self-reported data only
MIT researcher Justin Reich: comparing Alpha students to general population, not to similarly-resourced private school peers — comparison is misleading
$40-75K/year tuition: not a scalable equity solution
Northwestern University education researcher: "not supported by critical research" and "an open experiment"
No character or whole-person development framework published

Evidence: Tier 3 — Interesting Proof of Concept, Unverified The results are worth watching. The evidence does not yet meet an independent verification standard. This does not mean it isn't working — it means we don't know yet.

Approach 5

Spirit-Bridge

Full AI Learning Ecosystem • Derech Technologies

What the Architecture Demonstrates

Homeschool pricing (~$1,200/year per family) undercuts private tutoring by 75–90%. A skilled human tutor brings relationship depth and contextual judgment AI cannot fully replicate — but AI offers unlimited patience, 24/7 availability, and consistent personalization across multiple children that no individual tutor can match at this price point
The only platform in this comparison with all seven research-backed teaching principles enforced in the AI engine itself — not marketing copy
Vygotsky, Ebbinghaus, Hattie, Bloom, and Self-Determination Theory baked into code across AdaptiveLearningEngine, MasteryTracker, CurriculumTeachingEngine
CharacterGrowthTracker: 13-year longitudinal tracking of Fruit of the Spirit + Universal Virtues — no other platform attempts this
Safe local network deployment: structurally impossible for students to access social media or YouTube — not policy-dependent
Emotional Wellness is Priority #1: lesson stops before content if student is distressed
Original purpose-built curriculum across 8 subjects with a documented educational philosophy audit

Honest Limitations — Required by Our Own Standard

Pre-launch platform: features in this comparison reflect the complete system at release, not current development state. All identified features are committed for completion prior to launch.
No published peer-reviewed outcome data yet. The architecture is built on Tier 1 learning science. Platform-specific longitudinal results are emerging, not yet established.
Academic curriculum currently covers Grades 5–7; Life Readiness covers Grades 9–12. K–4 and Grade 8 academic curriculum expansion is planned. Readers should know this.
Disclosure: the author of this analysis is the developer of Spirit-Bridge. The Organizational Lens Principle applies to us as much as to anyone we analyze. We are saying so plainly because we believe credibility requires it.
As with any AI system: the design guards against misuse, but human deployment decisions introduce variables no design can fully control.

Evidence Tier: Tier 3 — Principled Architecture, Results Emerging The architecture is built around Tier 1 validated learning science. The comparison score reflects the complete designed system. Platform-specific longitudinal outcomes are not yet published — the same honest standard we apply to every platform on this list.

The Spirit-Bridge Architecture

What "Purposeful" Actually Looks Like — Built Into the Code

One of the recurring criticisms of this entire category is that "purposeful AI tutoring" is a marketing phrase, not a technical reality. The Chromebook rollout was also called purposeful. So was the 1990s multimedia classroom. So here is what purposeful looks like in Spirit-Bridge — not in a brochure, but in the actual system architecture.

The CurriculumTeachingEngine builds the AI prompt in four layers: the companion's personality, the learning track overlay (Biblical or Character), subject-specific pedagogy, and the lesson content. Embedded permanently in the third layer are seven teaching principles drawn from Hattie, Bloom, Vygotsky, Self-Determination Theory, and the testing effect — enforced in every session, with every student, regardless of the lesson.

Holding All of It Simultaneously

The truth is not simple and we won't pretend it is. Here is what the evidence actually supports — including the things that are inconvenient to say.

Verified — Tier 1

Educational technology has a genuine century-long failure record — radio (1937), TV (1950s), classroom computers (1980s), laptops (2000s), tablets (2010s). The pattern is not bad technology. It is the same assumption repeated: that access produces outcomes.

False — Tier 4

"Technology failed students." This is the headline. It is not what the researchers are saying. Even the neuroscientist cited most often by the anti-screen movement wrote: "This is not a debate about rejecting technology. It is a question of aligning educational tools with how human learning actually works."

Interpretation Required — Tier 2

Alpha School's top MAP test results are real data. Whether they prove the model works — or prove that $55,000/year resources, curated enrollment, and a motivated parent base produce outcomes — is a genuinely open question. Both explanations fit the available evidence.

Unverified but Directionally Strong — Tier 3

Spirit-Bridge's architecture is built around Tier 1 validated learning science. The platform-specific longitudinal outcomes are not yet available. "The design is right" and "the results are proven" are not the same claim. We are making the first claim, not the second.

What responsible adoption looks like: intentional, pedagogically grounded, teacher-supported, age-appropriate, transparent with parents, humble about what we don't yet know. The call to action is not "use more technology." It is: finally ask the question we should have been asking for a hundred years — not what tool, but what outcome, and how does this tool serve it?

The assumption was always wrong. Now we know it. The question is whether we use that knowledge to go backward — or forward.

The Comparison That Should Have Been Done Twenty Years Ago: EdTech Approaches, What Failed, and What's Actually Different Now

Standard Disclosures

Before We Compare Anything — We Need a Standard

What We're Comparing — and Why These Five

Traditional Public School + Standard EdTech

Adaptive Practice Platforms

Khan Academy / Khanmigo

Alpha School

Spirit-Bridge

The Pedagogical Comparison Matrix

Strengths, Weaknesses & Evidentiary Tier

Traditional Public School + Standard EdTech

Adaptive Practice Platforms

Khan Academy / Khanmigo

Alpha School

Spirit-Bridge

What "Purposeful" Actually Looks Like — Built Into the Code

The 7 Teaching Principles — Hardcoded Into Every Session

AdaptiveLearningEngine

MasteryTracker

InterventionSystem v4.0

CharacterGrowthTracker

GiftDiscoveryEngine

Safe Local Network

Holding All of It Simultaneously

A Note on Sources — Organizational Lens Applied

Sources