Under The Hood
In Your Hands
The Big Picture
Live Demo
90 seconds. A complete competency portrait.
Watch the engine score a student response in real time, with live signals.
See It Measure →
The Solutions
The Methods
The Foundations
Selected by California
Chosen To Design The State's Behavioral Health Performance Management System.
The BHSSA award — durable-skill measurement at state scale.
See The Work →
The Case
The Science
The Proof
The Proof
Real districts, real data, real names.
What partner districts saw in year one — in their own words.
See The Case Studies →
The Arc
The People
The Promise
From the Founders
We built this because the tools didn't exist.
A note on why two educators chose to build a measurement company.
Read Our Story →
NEXT GENERATION ASSESSMENT

Every Voice Carries Measurable Skills

Live · measuring

Skills, Quantified.

CAMILACalifornia · Grade 12
ERS—/100
Evidence richness
IMPACTER PATHWAY· HUMAN-MODEL κ 0.89· WEIGHTED κ 0.87· ADJACENT AGREEMENT 98.0%· EXACT AGREEMENT 77.3%· PAIRWISE AGREEMENT 91%· RATER DRIFT < 2%· TEST-RETEST r 0.86· SPLIT-HALF 0.88· INTERNAL CONSISTENCY α ≥ 0.85· INTRA-CLASS CORRELATION 0.87· EAP RELIABILITY 0.90· TLI 0.954· SRMR 0.038· χ²/df 2.14· FACTOR LOADINGS ≥ 0.65· CONVERGENT r 0.74 · DISCRIMINANT r 0.31· HTMT RATIO 0.42· CONSTRUCT VALIDITY r 0.74· CRITERION r 0.61· KNOWN-GROUPS d 0.83· SCALE 200–800 · μ 500 · σ 100· TRAINING n = 38,204· VALIDATION n = 9,551· HOLDOUT n = 4,102· 95% CI ±0.02· POWER 0.94 · α 0.05· EFFECT SIZE d = 0.71· D-STUDY OPTIMAL n_r = 2· RASCH SEPARATION 3.4· INFIT MNSQ 0.97· IMPACTER PATHWAY· HUMAN-MODEL κ 0.89· WEIGHTED κ 0.87· ADJACENT AGREEMENT 98.0%· EXACT AGREEMENT 77.3%· PAIRWISE AGREEMENT 91%· RATER DRIFT < 2%· TEST-RETEST r 0.86· SPLIT-HALF 0.88· INTERNAL CONSISTENCY α ≥ 0.85· INTRA-CLASS CORRELATION 0.87· EAP RELIABILITY 0.90· TLI 0.954· SRMR 0.038· χ²/df 2.14· FACTOR LOADINGS ≥ 0.65· CONVERGENT r 0.74 · DISCRIMINANT r 0.31· HTMT RATIO 0.42· CONSTRUCT VALIDITY r 0.74· CRITERION r 0.61· KNOWN-GROUPS d 0.83· SCALE 200–800 · μ 500 · σ 100· TRAINING n = 38,204· VALIDATION n = 9,551· HOLDOUT n = 4,102· 95% CI ±0.02· POWER 0.94 · α 0.05· EFFECT SIZE d = 0.71· D-STUDY OPTIMAL n_r = 2· RASCH SEPARATION 3.4· INFIT MNSQ 0.97·
Predictive Power for Adult Life Outcomes
Cognitive Skills Durable Skills
Hover bars for research citations
The Quantification Gap

What Matters Most
Goes Unmeasured

Durable skills are 2× more predictive of adult outcomes than cognitive metrics. Yet every tool the field has ever built to measure them has been incomplete. Until authentic voice.

Surveys.
An unreliable pulse.
While five choices can measure a moment in time, they often miss the patterns that produce it.
Reflections.
A small window into learning.
Written responses show the answer, but not the deep and critical thinking it took to get there.
Voice.
The architecture of becoming.
Two minutes of authentic language. Rubric-aligned scoring, validated by expert human raters at scale.
Methodology · 18 Meta-Analyses, 2006–2023 · N > 500,000 Students
Moffitt Weissbourd Duckworth Von Stumm
Student Spotlight · Deep Dive

One Response, Five Signals

Two minutes of authentic language, scored across five durable skills.
Press play and watch the signals land.

Julian · Grade 11
Independence High · Pleasantville ISD

Tap unmute to hear Julian speak.

↑ TAP TO HEAR JULIAN
0:00 / 2:03
Durable skills in this response 0 detected
Listening for durable skills
Five skills are scored from this one response. As Julian speaks, whichever skill his language triggers appears here — with the phrase that earned it, the rubric level behind the score, and a jump back to that moment.
Purpose1:22listening…3.8/4L4 · Strong
"My main goal in life is to be successful and make my family proud."
Rubric L4 · Goal Framing & Future OrientationValue Statement + Purpose Verb + Abstract Mission Phrase
Grit1:11listening…3.6/4L4 · Strong
"I even take on tests that I know are too big for me, but I end up doing it anyways."
Rubric L4 · Temporal Sequencing of EffortTask Difficulty Marker + Persistence Verb
Growth Mindset0:32listening…3.4/4L3 · Strong
"Growing up in that environment forces you to become a better version of yourself."
Rubric L3 · Learning Progression FramingEnvironmental Framing + Trajectory Verb
Perspective-Taking0:09listening…2.9/4L3 · Present
"Failure is defined in many different ways — people have their own definitions."
Rubric L3 · Perspective Modeling · Theory of MindAlternative Definition + Other-Attribution
Gratitude0:38listening…2.7/4L3 · Present
"One of my main motivators is definitely my auntie. She drives me to push myself."
Rubric L3 · Reciprocal Recognition · Named AdultNamed Adult + Attribution of Motivation
See rater's notes Every score has a source — rubric-aligned to Harvard's Making Caring Common project.
Our Measurement Core

Eight Durable Skills.
Identified By Harvard.
Predictive Of Success.

From Harvard's Making Caring Common project. Dr. Richard Weissbourd and his team identified these eight skills as the strongest predictors of success in college, career, and relationships. Every student already carries evidence of them. Click any skill to see its definition.

Curiosity Perspective-Taking Purpose Self-Control Grit Growth Mindset Compassion Gratitude
Definition
Click any skill above to see its definition.
↗ Definitions of Key Character and Non-Cognitive Terms and Skills Harvard
↗ Anchor Attributes — Wheel and Definitions Impacter Pathway
Prompt Science

What You Ask Determines
What You Can Measure

Open-ended and linguistically calibrated, each prompt is engineered to elicit language patterns that map directly to the Anchor Attributes.

Perspective-Taking

The Origins Of Your Style

Where students trace the influences that shaped them — a direct signal of relational awareness and identity formation.

6 SIGNALS
Growth Mindset

When You Bounced Back

How students narrate failure and recovery — whether they frame setback as a ceiling, or just a step in the process.

5 SIGNALS
Compassion

Someone Who Cares

The people students name when they feel truly seen, and what that language reveals about their capacity to connect.

5 SIGNALS
Grit

Something You're Working On

The effort students own when no one's watching, and the language that tells you they actually mean it.

6 SIGNALS
The Difference

Surveys Miss It
Voice Reveals It

A Likert scale compresses a student's entire motivational architecture into a single number.
Authentic voice preserves the human signals that make that motivation real and actionable.

✕  Traditional Survey

Rate your feelings 1 to 5.

Student Wellness Check-In 3 of 3
Q1. I am motivated to succeed in school.
SD
1
2
3
4
5
SA
Q2. I feel connected to people at my school.
SD
1
2
3
4
5
SA
Q3. I have access to support when I need it.
SD
1
2
3
4
5
SA
Student answered in ~8 seconds. No qualitative signal captured.

A score of "3" doesn't tell you why a student is motivated, or whether they're one tough week away from disengaging entirely.

✓  Authentic Voice

Tell us what drives you.

Click to see what was detected →
"So I moved here to Lead Elementary. And I'm glad that I moved here because the school has given me the possibility to learnBelonging Moment, not get bullied, and just feel myself and express myself from everything I've learned. And I am happy here. And I love the support that I get from hereLearning Joy instead of that old school I used to go to. I wish that all students can just get a chance to meet new teachersCommunity Empathy, have a nicer community, and just have friends, cause I used to never have friends."
1
BELONGING MOMENT · L4 · EXCEPTIONAL
"the school has given me the possibility to learn"
Student names the school as the direct enabler of growth — a clear before/after arc. Self-expression as the core benefit of belonging.
2
LEARNING JOY · L3 · CLEAR
"And I am happy here. And I love the support that I get from here"
Sustained positive affect tied directly to school environment. Two belonging signals in one statement.
3
COMMUNITY EMPATHY · L3 · CLEAR
"I wish that all students can just get a chance to meet new teachers"
Student generalizes her own experience outward — wishing belonging for all peers. Strong indicator of social belonging and perspective-taking.
← Click to flip back

Trusted Across 5 States And 200+ Schools

What's Next

Which Skills Can
We Measure For You?

Bring us your students, your framework, or your district. We'll build the portrait.

Path 01
Bring Your Portrait Of A Learner
We capture authentic student voice against your Portrait of a Graduate framework or whatever durable skills your community has already defined.
Try For Yourself →
Path 02
Crosswalk Your Framework
Run your district's existing framework — LCAP, MTSS, behavioral health — through our Crosswalk Engine to see exactly where voice maps to what you already track.
See the Crosswalk Engine →
⤢
Provenance · How This Was Measured

Composite Portrait · 3.4/4.0

Model & Method
DeBERTa-v3 + CORAL ordinal regression, rubric-aligned to the IMPACTER 8-attribute framework. Model Suite v6.0. Every metric corresponds to a specific rubric criterion (Level 1–4). Human raters set the standard — the model is calibrated against their inter-annotator agreement.
Training & Validation
Training corpus6.2 billion words
Rubric-scored responses38,204
Holdout set (never seen by model)n = 4,102
Credentialed evaluators, calibrated quarterly47
Agreement with Human Raters
Human–Model κ (kappa)0.89
Human–Human κ (baseline)0.87
Quadratic Weighted Kappa0.918
Exact agreement77.3%
Adjacent agreement (± 1 level)98.0%
Reproducibility
Score variance (5 identical runs)0.02
Test–retest reliability r0.86
Internal consistency α≥ 0.85
Framework Crosswalk
Validated against the CDE Portrait framework, CASEL 5 competencies, and Harvard's Making Caring Common research attributes. Framework-agnostic scoring — ingest any competency set, deliver consistent rubric-aligned data.
Rater's Notes · Julian · Grade 11 · Independence High
Response · 2:03 · scored 07/03/2026 · click any circle to play from that mark
0:00 / 2:03
solid response
Notes reviewed against IMPACTER 8-attribute rubric, Model Suite v6.0. Scoring agrees with model output within ± 1 level on all 5 attributes. — RM