The Learning Cognitive Profile.

Every learner is profiled before instruction begins. The Learning Cognitive Profile (LCP) is a 30-question AI-generated assessment delivered in three batches of ten, structured as five named categories with a fixed distribution and measured against four canonical learning-style dimensions. The measurement model is invariant; the presentation adapts to the learner. The resulting profile drives every lesson the platform generates for that learner thereafter.

30 questions across 5 categories. The distribution is not a recommendation.

• Cognitive Preference Probe

  12

  The structured anchor — three questions per dimension. The measurement model is invariant; the wording adapts to age and language.

• Scenario-Based Decision Simulation

  6

  A short scenario, two viable paths, the learner picks. Reveals reasoning preference without asking the learner to introspect about it.

• Emotional-Cognitive Resonance Rating

  6

  How a learner responds to ambiguity, novelty, repetition. Calibrates the affective dimension that downstream content must navigate.

• Meta-Cognitive Reflection Prompt

  4

  The learner reflects on a recent learning experience. Tests self-model accuracy — a stronger signal than direct self-report.

• Behavioral Divergence Trigger

  2

  Two questions designed to surface counterexamples. The dimensions you self-report and the dimensions you behave under can differ.

What the LCP actually measures.

• Visual ↔ Verbal

  Whether the learner reasons better through diagrams, hierarchies, and spatial structure — or through prose, definition, and sequence in language.

• Active ↔ Reflective

  Whether the learner consolidates by doing and discussing — or by thinking and re-reading. Felder-Silverman, two-poled, strength-rated.

• Sensing ↔ Intuitive

  Whether the learner anchors on concrete observation and applied examples — or on patterns, abstractions, and theoretical scaffolds.

• Sequential ↔ Global

  Whether the learner builds understanding step-by-step, with each piece resting on the one before it — or in larger leaps, holding context until the structure clicks.

The dimensions follow the Felder-Silverman tradition. Each is two-poled, strength-rated. The result is a four-coordinate signature that downstream content generation reads structurally — not a personality label.

What the generation prompt enforces.

1. 01

   30 questions across 5 categories with a fixed distribution: Cognitive Preference Probe (12), Scenario-Based Decision Simulation (6), Emotional-Cognitive Resonance Rating (6), Meta-Cognitive Reflection Prompt (4), Behavioral Divergence Trigger (2).

2. 02

   4 measurement dimensions, two-poled (Felder-Silverman tradition): visual–verbal, active–reflective, sensing–intuitive, sequential–global.

3. 03

   Generation system prompt is explicit: number of questions, category distribution, and measured dimensions cannot be altered. Only the presentation (wording, scenario context) adapts to the learner’s age, language, and institutional setting.

4. 04

   Profile parser produces a per-dimension direction + strength score; the strength is exposed to downstream content generation via a structured formatLcpForPrompt() utility.

5. 05

   Output: a profile summary blob (Markdown) and a structured responses blob; both are tenant-scoped and FERPA-treated as education records.

Why the measurement model cannot adapt.

Most "learning style quizzes" in market today flatter the learner. The questions move toward the answers the learner appears to like. The model is doing user research, not measurement.

The LCP does the opposite. The category distribution is fixed (12 + 6 + 6 + 4 + 2). The four dimensions are fixed. The instruction to the generation model is explicit: do not change the number of questions, do not modify the dimensions being measured, do not introduce bias or leading phrasing, do not personalise content in a way that alters the measurement model. The only thing that adapts is the surface — vocabulary, scenario context, locale references — so a 9-year-old in Lahore and a 50-year-old in Sacramento can both answer the same instrument honestly.

That is what makes the LCP a profile worth reading. And it is the substrate that every personalised lesson the platform generates afterward is built on.