The Educational Performance Tension Index (EPTI)

1 Educational Performance Tension Index (EPTI) A Non-Weighted Structural Model for Outcome-Based Learning Diagnostics Author: Jim Y. Huang*1 Executive Summary Most educational performance indicators are constructed through pre-assigned dimensional weights. Coursework may count for 25%, examinations for 40%, participation for 10%, and socio-economic modifiers may or may not enter the equation. These allocations are rarely structural — they are judgments made before the system is observed. The metric is defined first; reality is interpreted second. EPTI reverses the logic. Instead of deciding importance in advance, importance emerges from observable tension in the outcome gradient. A module matters only if its disturbance is reflected in performance instability. No valuation. No normative scoring. No assumed weights. EPTI offers a reproducible mathematical method to translate student performance data into tension surfaces along a single performance axis, identifying friction zones in the learning process without asserting causation or blame. It does not predict scores. It reveals where a system is absorbing friction. This whitepaper introduces the conceptual foundation, axioms, structural architecture, and mathematical formulation of EPTI as a non-subjective diagnostic framework for educational systems. 1. Background Educational evaluation traditionally follows this pattern: 1. Choose dimensions (assignments, tests, attendance, etc.) 2. Assign subjective weights * Doctoral Researcher , University of Toronto (OISE). Chartered Professional Accountant (CPA, Canada), Cer�fied Prac�sing Accountant (CPA, Australia), and Trust and Estate Prac��oner (TEP). Research focuses on fiscal architecture, cross-border capital movement, intergenera�onal post-tax capacity, and ins�tu�onal inequality. Contact email : jimy[email protected]oronto.ca 2 3. Aggregate into a score This architecture carries three systemic limitations: • Weights reflect expert preference, not structural necessity • Comparability collapses over time or across regions • Performance becomes interpreted, not measured When curriculum changes, policies shift, or demographic patterns evolve, the weight matrix lacks portability. A structural metric must not rely on subjective coefficients. EPTI reframes performance as the processed output of sequential learning modules, where tension is detected by deviations in expected outcome behavior. The result becomes evidence; friction becomes information. 2. Structural Positioning EPTI belongs to a broader family of tension-based visibility tools: Metric Object Geometry Input Visibility Mechanism ITI Institutions XY-plane Fiscal movement Structural tension IDI Integrity Arithmetic vs geometric Rule vs behavior Distortion gap EDI Emotion Affect field Interaction events Sentiment tension EPTI Education Single performance axis Outcome vectors Performance tension No hierarchy. No replacement. Each renders a different layer of reality visible. 3. Axioms Axiom 1 — Outcome Primacy Only outcomes are measurable. Effort, intention, and pedagogy are not metrics unless manifest. Axiom 2 — Tension Reveals Structure Where performance deviates from expected trend, friction exists. Deviation is not failure — it is information. Axiom 3 — Weights Must Be Emergent Importance is not assigned. Importance is inferred from outcome sensitivity. 3 Formally: Importance(𝑀𝑀𝑖𝑖) =∣∂𝑌𝑌 ∂𝑀𝑀𝑖𝑖∣ where 𝑀𝑀𝑖𝑖is a module and 𝑌𝑌is outcome. Together, these axioms define EPTI as a reverse-solved, not pre-weighted model. 4. Model Architecture Learning is represented as a processing spindle — a performance axis along which modules interact with the learner. Modules act as cutting tools; the axis rotates forward. The learner is material undergoing transformation. Friction accumulates where the cut is unstable. [Input → Learning Axis → M1 → M2 → M3 → … → Output] ↑ Tension Zones • Modules are not inherently heavier or lighter • Friction determines impact, not design • Tension = resistance observable in outcomes Unlike scoring models, EPTI does not evaluate components. It detects zones where performance begins to strain. 5. Mathematical Formulation Let: • 𝑌𝑌𝑡𝑡= performance vector over time/cohorts • 𝑀𝑀1...𝑀𝑀𝑘𝑘= learning modules • 𝐸𝐸= unexplained residual (latent tension) Define module tension contribution: 𝑇𝑇𝑖𝑖=∣∂𝑌𝑌 ∂𝑀𝑀𝑖𝑖∣ Total tension: 4 𝐸𝐸𝐸𝐸𝑇𝑇𝐸𝐸 =�𝑇𝑇𝑖𝑖 𝑘𝑘 𝑖𝑖=1 +∣𝐸𝐸 ∣ Residual 𝐸𝐸is not noise — it signals unknowledged structural load. Weights never appear as inputs. They emerge from gradients. 6. Interpretation EPTI does not state: • who is responsible • which policy is good or bad • whether students tried hard EPTI renders friction visible, answering three questions: 1. Where in the process is tension concentrated? 2. Which modules correlate with performance variance? 3. How does friction evolve longitudinally? EPTI detects “Where,” not “Why.” Why belongs to intervention. Where belongs to measurement. 7. Applications At population scale, EPTI enables: • Province-wide performance tension maps • School vs school comparative diagnostics • Pre-/post-policy tension evaluation • Module-level friction visualization • Early warning for systemic decline • Evidence-based resource allocation Future reporting language could resemble: “EPTI indicates rising tension in the literacy module (+12%). Resource density suggests insufficient support capacity.” 5 No opinion. No blame. Only structural visibility. 8. Conclusion EPTI is not a scoring schema — scores are its raw input. EPTI is not an evaluative instrument — evaluation is political. EPTI is not a belief about learning — it is a method of reading tension. EPTI is the structural language of educational performance. When weights cease to be assumed and begin to emerge from the data itself, learning becomes measurable, governable, and legitimately comparable across time and jurisdiction. Appendix — Implementations Path (Non-Public Notes) 1. Build provincial EPTI baseline using EQAO / standardized outcome datasets 2. Produce first public diagnostic report for Ontario 3. Visualization set: heatmap, module tension profile, slope decay chart 4. Submit preview draft to Marvin (peer-review narrative) 5. Scale to Ministry, boards, media, research institutes 6. Licensing, recurring index subscription, API data integration