Four steps from session to insight
CLARO's pipeline turns Socratic dialogue into structured data that flows directly into your existing assessment ecosystem.
Socratic Session
Student engages in a DialogIQ™ session — no answers given, only questions asked.
Live Signal Capture
CLARO captures response latency, revision patterns, and reasoning depth in real time.
DQ + CFI Scoring
Two metrics computed: Dialogue Quality (reasoning depth) and Cognitive Fluency Index (processing efficiency).
Assessment Join
CLARO scores join to your i-Ready or MAP Growth data via Student ID through Clever.
* DQ and CFI are computed per session and aggregated across the school year into a student-level reasoning profile.
DQ and CFI — what they measure
Two complementary scores that together reveal the gap between what a student knows and how fluidly they can reason with it.
Reasoning Depth Score
Measures the cognitive complexity of a student's responses — not fluency or correctness, but the quality of reasoning demonstrated in dialogue.
- Captures reasoning quality, not just answer correctness
- Sensitive to growth that standardized tests miss
- Aggregates across sessions into a reliable yearly arc
Processing Efficiency Score
Measures how efficiently a student accesses and applies their knowledge — the difference between knowing something and being able to use it under pressure.
- Identifies students who know more than they show on tests
- Detects language suppression in multilingual learners (LSCS flag)
- Tracks scaffolding dependency over time — fading = growth
Five student patterns your assessments can't see
When you plot DQ against standardized assessment scores, five distinct patterns emerge — each with a different instructional implication.
Low DQ + low assessment score. Student is struggling with both reasoning process and content knowledge. Tier 2/3 intervention needed.
High DQ + low assessment score. Student reasons well but underperforms on tests — often multilingual learners with LSCS flag. Reduce remediation; increase dialogue challenge.
Low DQ + high assessment score. Student scores well but relies on memorization over reasoning. Shift to open-ended, reasoning-based tasks.
High DQ + high assessment score. Reasoning and content knowledge are aligned. Introduce peer facilitation and advanced challenges.
Mid-range on both dimensions. Monitor DQ arc across sessions; adjust scaffolding per domain subscore gaps.
Reliability: DQ α = 0.84 (n = 1,240 sessions) · CFI test-retest r = 0.79 (4-week interval)
40 years of evidence
CLARO's approach is grounded in the most replicated findings in educational research.
The 2 Sigma Problem
Bloom's landmark finding: one-to-one tutoring produces a 2 standard deviation gain — the most powerful intervention in education.
The Tutoring Effect
VanLehn's meta-analysis confirms: human and intelligent tutoring systems that require student reasoning outperform all other interventions.
Socratic AI Gains
MIT RAISE and independent K–12 studies find large effect sizes (η²=.20–.32) for Socratic AI tutors that withhold answers and return cognitive load to learners.
Bloom (1984) · VanLehn (2011) · MIT RAISE (2025) · Research Square K–12 Socratic AI (2025)
Works with what you already run
No migration, no parallel testing. CLARO reads the exports your district already produces.
Every pattern maps to a tier and a move
The gap analysis doesn't just label students — it hands coaches a specific instructional response tied to your MTSS framework.
"The average student under one-to-one mastery tutoring outperforms 98% of students under conventional instruction."Bloom (1984) · the "2 sigma" finding
CLARO's contribution isn't the tutoring effect itself — it's making the reasoning behind it visible and measurable, session by session.
See it with your district's own data
CLARO is an approved High Dosage Tutoring vendor for Fresno Unified School District, and works with any district running i-Ready or NWEA MAP Growth. Book a demonstration built on a sample of your data.
Bloom (1984). Educational Researcher, 13(6), 4–16. · VanLehn (2011). Educational Psychologist, 46(4), 197–221. · Kraft et al. (2021). Journal of Policy Analysis and Management, 40(3), 786–819. · MIT RAISE (2025). · Research Square K–12 Socratic AI (2025).
claroedu.com · draubrey@claroedu.com · Approved FUSD High Dosage Tutoring Vendor