| contato@retarazao.com
Login
Reta Razão
Ir para o conteúdo
Reta Razão
Cognitive Efficiency Engineering
Methodological showcase
May 2026
Cognitive Efficiency Engineering

Clinical diagnosis applied to STEM learning.

The shortest path between the problem and genuine understanding, with the fewest unnecessary detours.

Robert At Ramiarina
M.Sc. Biomedical Engineering · COPPE/UFRJ · Founder, Reta Razão
I The Problem

In an ordinary classroom, the bus stops near home — but everyone gets off at the same place.

The lesson plan is built for the group. The route never changes to accommodate individual reasoning. The hard work of figuring out what is missing falls entirely on the student — alone, in the bus seat, with the teacher up front, driving toward a different destination.

What is missing is not the content. It is the diagnosis of how each student actually thinks.

An error is not a sign of incompetence. It is a technical signal — it carries information about which reasoning step failed, at which point in the cognitive structure, with which pattern of recurrence. When that signal is read with rigor, the teaching that follows does not have to guess. It operates on data.

That is what diagnostic engineering delivers: an operating system for the continuous reading of a student’s reasoning — correct answers included — showing which paths are efficient, which carry hidden misconceptions, which need consolidation, and which can be advanced.

II The Discipline

Diagnostic engineering is not new. It is new applied to cognition.

The method carries a 24-year development trajectory. It was born in industrial engineering, adapted to hospital systems in peer-reviewed research, and since 2014 it has been applied to STEM learning. Each stage added rigor that survives into the next.

1997 — 2008
Technical Systems
Broadcast (TV Globo · 1998 World Cup), clinical engineering in a public hospital (1,200 devices, 200 beds), infrastructure at the Ministry of Health. FMEA applied to real systems.
2002 — 2008
Academic Research
M.Sc. COPPE/UFRJ. Publications in International Journal of Health Planning and Management (2007) and SciELO Revista de Saúde Pública (2008): diagnostic classification for resource prediction.
2014 — now
STEM Cognition
The same discipline applied to learning. A taxonomy of cognitive failure modes, operational metrics (ICS, IVR, RPN), and per-student diagnostic reports.
III The Method

Six traceable steps, from error to recommendation.

The FMEA protocol applied to cognition operates in a fixed sequence. Each step produces output for the next. Every output is auditable. The analysis has a verifiable beginning, middle, and end — it is not subjective interpretation, it is instrumentation.

Failure mode identification

Classification by the cognitive failure-mode taxonomy

Each error is categorized into one of the dictionary’s 12 families (40 subtypes). The classification answers: what is the cognitive type of the failure, not merely the topic where it appeared.

Root-cause analysis

5-Whys protocol applied to reasoning

Each mode is decomposed down to the operational root. It is not enough to know the error is one of “partial reading” — one must know which missing protocol produced the partial reading.

Consolidation Index per Skill (ICS)

Mastery metric over a discrete skill

Quantifies the degree of consolidation of each skill — fluent correct answers, hesitant correct answers, lucky guesses, errors. It ranges from 0 to 1. ICS < 0.5 indicates insufficient consolidation; ICS > 0.85 indicates operational mastery.

Resolution Speed Index (IVR)

Ratio of time spent to reference time

Measures temporal efficiency. IVR < 1.0 indicates fluent resolution; IVR > 1.5 indicates laborious reasoning or an inadequate strategy. Combined with right/wrong, it separates “knows it but is slow” from “does not know it and guessed”.

Risk Priority Number (RPN)

Weighted composite for prioritization

RPN combines the severity of the mode, the observed frequency, and the current detectability. It allows failure modes to be ranked by intervention priority — not everything is treated equally; the highest RPN becomes a dedicated lesson.

Operational recommendation

Executable output for the next teaching cycle

Each mode with a critical RPN generates a concrete pedagogical action: a dedicated lesson, question refraction, a verification protocol, a calibration exercise. The recommendation is specific, not generic.

Operational glossary

Technical vocabulary

FMEA
Failure Mode and Effects Analysis. Originating in industrial engineering, adapted in this methodology for the analysis of cognitive failure modes.
ICS — Consolidation Index
Quantifies mastery over a skill. Combines fluent, hesitant, and lucky correct answers and errors on a 0–1 scale.
IVR — Resolution Speed
Ratio of time spent to reference time. Detects laborious reasoning versus genuine fluency.
RPN — Risk Priority Number
A quantitative composite for ranking failure modes by pedagogical intervention priority.
TARC — True Accuracy Rate
A mastery metric that weights raw accuracy by fluency and the absence of luck. Applied in the domain calculator.
Failure mode
A recurrent cognitive pattern identified by operational taxonomy. It is not the specific error, but the class to which it belongs.
IV Application
Composite illustrative case

Student X — a synthetic profile built from real patterns.

A methodological composite of patterns observed across multiple students. No individual data is identifiable. The presentation follows the structure of the real reports delivered to families: first the reading of who the student is — in five movements — then the technical evidence that supports each move of the reading.

Profile

Age
16 years
Context
International school, bilingual curriculum
Initial diagnostic
Full SAT practice test (Bluebook)
Plan window
5 months, ~20 weeks

Score

1050
Baseline
1400+
Target

Initial composition: Math 510 · R/W 540. Realistic projection discussed below.

Student X in five movements

Before the numbers, the closest reading of Student X as a person, in five movements — mirror, surprise, mechanism, evidence, horizon.

I ·mirror
Student X is the kind of student parents recognize at home before any diagnostic: converses easily, holds formed opinions about what he reads, moves across subjects with genuine curiosity. At the dinner table, he explains concepts to a younger sibling. In class, he participates, asks questions, reads the assigned material. The report card says “engaged”. When the score arrives — 1050 — the first impulse is to treat it as carelessness, a bad day, anxiety. But the diagnostic reveals something else: Student X is reasoning the entire time, and losing precisely the one move the SAT is calibrated to detect.
II ·surprise
Where one would expect weakness, Student X is strong. On statistical probability questions — where students typically struggle — he scored 84%. On advanced punctuation, 88%; on sentence structure, 76%. In reading layers that normally require advanced textual maturity, he operates with ease. This is not a below-average student — it is a student with an unusual profile: strong precisely where many with higher scores are fragile.
III ·mechanism
And then, where he fails. Student X reads the problem, fires the right reasoning, but loses the link between the data read and the calculation executed somewhere along the way. In a linear equation, he sees he must isolate the variable, sets up the correct strategy, and on the last line flips the sign. In probability, he grasps the relationship, identifies the conditional event, and in the final fraction inverts numerator and denominator. He is not losing content — he is losing the verification point between steps that other students hold as a silent automatism. He is a student who operates on reliable intuition where he would need to operate by protocol.
IV ·evidence
On five linear-algebra questions that open the second module, Student X got all five wrong by the same move — what we internally call the dissociative mode: correct calculation, link lost on the last step. On three inference questions in Reading, the same trap in a different variant: he reads the prompt correctly, anticipates the answer correctly, and chooses the option whose wording most sounds like his anticipation — not the one the text actually supports. The test itself captures, with surgical precision, the student who operates on reliable intuition where he would need to operate by protocol. It is neither a linguistic vulnerability nor a conceptual weakness — it is a vulnerability to the design of the SAT distractor.
V ·horizon
The profile defines the route of the intervention. Student X does not need more content. He needs two specific fronts: a data-to-result verification protocol between calculation steps (which neutralizes the dissociative mode and consolidation inversion) and explicit training in evidence-versus-intuition discrimination in Reading (which neutralizes the self-imposed restriction of his repertoire). The other modes — overload on long prompts, pacing collapse in the final third — emerge as secondary, with standard treatment. Student X already has the engine; what he lacks is the instrument panel.
Profile in one sentence
“Student X operates on reliable intuition where the SAT demands a verifiable protocol.”
Structural Diagnostic · Student X

Failure modes identified — the evidence behind the reading

The narrative reading above is not subjective interpretation. Each move described maps to a quantified failure mode, ranked by priority.

ModeOperational descriptionICSIVRRPN
Dissociative mode
Right setup, wrong execution
Reads the prompt, fires the correct procedure, but loses the link between the data read and the calculation executed. The error appears in execution, not in comprehension.0.341.42428
Consolidation inversion
Advanced mastered, fundamentals gapped
Apparent mastery in advanced topics, gaps in fundamentals. The classic pattern of an international curriculum that skips consolidation.0.411.18385
Textual overload
Drop on dense prompts
Performance drop proportional to prompt length. It is not a lack of vocabulary — it is the absence of a segmented-reading protocol.0.521.68312
Self-imposed restriction
Fixation on the first strategy
The student fixes on an initial strategy and does not test alternatives, even when the result contradicts it. Arbitrarily narrows the scope of reasoning.0.482.04298
Pacing collapse
Drop in the last 10 minutes
The accuracy curve drops abruptly in the final third of the test. It is not fatigue — it is the absence of prior pacing and a skip strategy.0.630.92186

Domain Map — TARC calculator by area

Each cell represents an assessed sub-area. The color indicates the degree of operational consolidation. Red areas are absolute intervention priorities; green areas are the student’s cognitive assets and must be preserved.

Heart of Algebra0.31
Linear Systems0.42
Functions0.58
Geometry0.61
Statistics0.72
Probability0.84
Word Problems0.39
Data Analysis0.55
Inference0.28
Evidence Cmd0.44
Vocabulary0.60
Structure0.71
Rhetoric0.54
Standard Eng0.76
Punctuation0.88
Sentence Form0.79

Operational recommendation — derived from the critical RPNs

  1. Data-to-result verification protocol. For the dissociative mode (RPN 428), introduce a mandatory coherence check between the data read and the calculation executed, before marking an answer. Applied in the first 4 Math sessions, it becomes automatic by week 6.
  2. Consolidate fundamentals before advanced topics. For consolidation inversion (RPN 385), suspend exercises on topics where ICS is low and return to Heart of Algebra and Inference — areas with critical TARC in the domain map. Without consolidation here, advancing only amplifies the fragility.
  3. Segmented reading with anchors. For textual overload (RPN 312), introduce a reading protocol in 2-sentence blocks with re-reading anchors. Applied in Reading exercises from the first week of R/W.
  4. Strategic divergence training. For the self-imposed restriction (RPN 298), exercises with the explicit instruction “solve it two ways”. Breaks the fixation on the first strategy chosen.
  5. Pacing and skip protocol. For pacing collapse (RPN 186), strategic-skip training on full practice tests from week 5 of the Sprint.
The immediate potential of this diagnostic

What Student X recovers from a careful reading of this report alone

Even before the first conceptual remediation session, three takeaways that follow directly from this reading produce measurable, auditable gains:

  • Recognition of the five trap families — the student begins to identify the shape of the distractor before marking. The effect is not uniform (it depends on triggering the protocol under time pressure), but for 2 to 3 additional Math questions per test, explicit familiarity is enough.
    +30 to +40 pts · Math
  • Data-to-result verification protocol — between an intermediate step and the final answer, the student interrupts the silent automatism and checks coherence. It directly reverses the dissociative mode in cases where the calculation was correct and only the last step flipped.
    +20 to +30 pts · Math
  • Evidence-versus-intuition discrimination discipline — in Reading, the student begins to ask “is this option supported by the text, or does it just resonate with my anticipation?” before marking. It directly neutralizes the self-imposed restriction of his repertoire.
    +30 to +40 pts · R/W
+80 to +110 points
Projected SAT total: from 1050 to 1130–1160, from internalizing what this diagnostic already delivers alone — before any conceptual remediation session.

Realistic projection — with a plan tailored over the RPNs

Based on the typical absorption curve observed when the intervention is ranked by RPN, with a diagnostic review every 4 weeks. It starts from the score already internalized via the report (previous item), not from 1050.

Week 8
1240
Consolidation of Heart of Algebra & Inference. Dissociative and inversion modes under control.
Week 14
1340
Foundations complete. Sprint underway. Textual overload and self-imposed restriction in an elimination routine.
Week 20
1420+
Sprint complete. Pacing automated. End-of-test collapse eliminated. Margin for test day.
V From Analysis to Course

The diagnostic is not a loose analysis. It is the input that defines every lesson that follows.

Each of Student X’s critical RPNs demands a specific lesson, with tailored material. It is not a selection of pre-existing exercises — it is production. The same failure mode in another student generates a different lesson, because the root cause is different. Below, an example of a lesson built over the dissociative mode in an IB Biology context.

IB MYP 4 Biology — Criterion C: Processing & Evaluating

Catalase × Temperature · Lesson built over the dissociative mode + data-to-result verification protocol
[Screenshots or short demo video of the course — to be integrated]
Suggested: 3-4 screenshots of real lessons from the IB Criterion C Biology course,
or a 60-90 second video walking through a complete lesson.

The lesson presents the classic catalase-versus-temperature experiment, but with explicit pedagogical instrumentation: mandatory verification points between the data collected and the conclusion drawn. The student cannot advance without closing the coherence checkpoint. The design is diagnostic — the lesson itself works against the dissociative mode.

VI Who Operates

Twenty-five years of the same discipline, applied to three domains.

Biomedical engineer (M.Sc. COPPE/UFRJ, 2009), with a trajectory that began in broadcast telecommunications (transmission engineering for the 1998 World Cup, TV Globo) and moved to clinical engineering in a public hospital (Hospital da Lagoa, 1,200 devices, 200 beds, 2000-2008). At the Ministry of Health (2004-2008), he coordinated federal clinical infrastructure and saw his work recognized with a first-place finish in a national selection.

The research at COPPE/UFRJ applied hospital administrative data and ICD-10 comorbidity adjustment to the classification of clinical states and resource prediction — diagnostic engineering for resource optimization. Published in International Journal of Health Planning and Management (2007) and SciELO Revista de Saúde Pública (2008). The origin of the current diagnostic method dates to the CBEB 2002 conference.

Since 2014, the same discipline applied to STEM learning. He developed the cognitive failure-mode taxonomy, the domain calculator, and the cognitive FMEA protocol now consolidated in the Reta Razão platform. Author of Math Bridge: A Bilingual Dictionary of Mathematical Terms in English and Portuguese (ISBN 978-65-01-29032-4). Students currently at Cambridge, Stanford, Cornell, Duke, UCLA, UCL.

The guiding question runs across all three decades: what makes a biological system truly efficient? The domains change — broadcast signals, hospital flows, cognitive reasoning. The question does not.

Education

  • M.Sc. Biomedical Engineering — COPPE/UFRJ
  • Licentiate in Mathematics — UNESA
  • Electronics Engineering — CEFET-RJ

Publications & Recognition

  • IJHPM 2007 · SciELO RSP 2008
  • CBEB 2002 (origin of the method)
  • Math Bridge (2024)
  • TCU Ruling 2911/2016 · 3 first-place finishes in national selection
Next Step

If your family, school, or institution needs diagnostic engineering for STEM learning — let’s talk.

The first step is a 30-minute demo session, at no cost. I present the method applied to the specific profile of the student or context, show the Domain Map live, and answer questions. It is not a sales pitch — it is a technical reading of what makes sense for your case.

Available in person in Leblon, or by international video call.

Reta Razão
Av. Ataulfo de Paiva, 1174B · Leblon · Rio de Janeiro
Rolar para cima

Este site utiliza cookies essenciais para funcionamento e rastreamento de conversões de anúncios. Saiba mais em nossa Política de Cookies.