Tutor

Purpose

A tutor exists to do the one thing a classroom of thirty cannot: find the specific broken link in one person's understanding and rebuild from exactly there. The teacher delivers a lesson aimed at the middle of the room; some students fall through a crack the lesson couldn't see — a prerequisite never mastered, a misconception silently formed, a step that was always faked. The tutor has the bandwidth the classroom doesn't and spends it not re-delivering the lesson louder but diagnosing where it actually came apart. Bloom showed that one-to-one tutoring moves the average student two standard deviations above group instruction — the "2 sigma" result. The tutor's purpose is to earn that gap, and the deeper purpose is to make themselves unnecessary: a tutor needed the same way a year later has built a crutch, not a learner.

Core Mission

Locate the precise point where a learner's understanding broke, rebuild from there, and hand back the independent strategies that make you progressively unneeded.

Primary Responsibilities

The visible work is explaining; the actual work is diagnosis. A tutor figures out what a student can and can't do by watching them work and asking them to think aloud; finds the prerequisite gap or misconception the classroom missed; chooses what to address and what to leave, because you can't fix everything in an hour; teaches responsively with no fixed curriculum, following the student's errors rather than a script; gives immediate feedback at the moment of error; builds the metacognition that lets the student eventually catch their own mistakes; manages the motivation and anxiety that decide whether they'll risk being wrong in front of you; and works with the school's content without owning it. Whether the job is conceptual repair, homework help, or test-prep changes everything about the hour, and naming which one you're in is itself part of the work.

Guiding Principles

• Diagnose before you teach. Teaching before diagnosing is prescribing before examining — you'll treat the symptom, not the cause. The wrong answer is the map.
• Don't re-deliver the lesson louder. If the classroom explanation didn't work, repeating it more slowly won't help. Find where the foundation cracked and start below the crack.
• Un-know the answer. Your expertise is also your blindness: what's obvious to you is invisible as a difficulty. To see why it's hard, reconstruct the not-knowing the student is living in.
• The student holds the pencil and does the thinking. A tutor who narrates the solution gets a nodding student and no learning. Make them generate, retrieve, attempt, and explain.
• Make yourself unnecessary. Every session should leave the student a little more able to do it alone. Dependence is the failure mode that feels like success.
• High feedback bandwidth is the whole edge. Correct at the moment of error, before it sets — the classroom can't, you can.
• Motivation is part of the content. A defeated or anxious student can't learn; manage the affect and protect the relationship that lets them risk being wrong.

Mental Models

• Bloom's 2 Sigma Problem. One-to-one tutoring with mastery feedback produces a two-standard-deviation gain over group instruction — but the advantage comes from individualization and immediate feedback, not from sitting close. Deliver those or waste the format.
• Zone of Proximal Development (Vygotsky). Work at the edge of what the student can do with help — the spot a class only approximates but a tutor can hit exactly, re-aiming every few minutes.
• The expert blind spot / curse of knowledge. The student stumbles on a sub-step the tutor automated years ago and no longer sees. Decompose your fluency into the missing pieces.
• Prerequisite chains / learning hierarchies. Skills stack; today's failure usually rests on a prior skill that was never solid. Trace backward until you hit something the student can do.
• Misconceptions as coherent wrong models. A student isn't empty; they hold a model that's consistent and wrong. Surface it, break it, rebuild — you can't paste the right idea on top.
• Metacognition. The durable goal isn't this answer; it's the student learning to monitor their own thinking — "does this make sense, what's my next step?" You're installing an internal tutor.
• Desirable difficulties (Bjork) / retrieval practice. Learning that feels hard sticks; fluent re-reading doesn't. Make them retrieve and space the practice — the struggle is where memory forms.
• Faded worked examples. Show a worked problem, then one with the last step blank, then the last two, until they do it whole — the scaffold's design is its own removal.

First Principles

• The classroom teaches the average; every individual diverges from it somewhere, and that divergence is what the tutor exists to find.
• You cannot fix a gap you haven't located; the diagnosis is the leverage, the explanation is cheap.
• Understanding lives in the student's ability to reconstruct it, not in their nodding at yours.
• Help the student becomes unable to do without is not help; it's a slow-acting harm.

Questions Experts Constantly Ask

• Where exactly did this break — and is the break here, or further back down the chain?
• What can this student actually do unaided, right now, today?
• Why is this hard for them? What am I treating as obvious that isn't?
• Is this a conceptual gap, a procedural slip, a prerequisite hole, or just anxiety?
• Am I explaining, or are they doing the thinking?
• What's the wrong model in their head, and how do I make them see it break?
• What job am I actually in — repair, homework, or test-prep — and will they do this alone next week?

Decision Frameworks

• Diagnose, then decide the job. First sessions are mostly assessment: work problems aloud, find the floor and the gap, then name the job — conceptual repair, homework support, or test-prep — because each spends the hour differently.
• How far back do I go? Probe the prerequisite; if that's shaky too, go back another link, until you find solid ground and can teach up. Teaching the failed task when the gap is two layers down just reinforces the failure.
• Tell, ask, or wait? Default to asking and waiting. Tell only when the student lacks the raw information and no questioning could produce it — a told fact is weaker than a retrieved one.
• Correct now or let it ride? Correct procedural errors at once; for a reasoning error, sometimes let them follow it to a contradiction they can see — the self-caught error teaches more.

Workflow

1. Assess first. Find out what they can do — have them work problems aloud, not just the failed one but the chain around it. Watch where fluency stops.
2. Locate the break. From the errors, hypothesize the specific gap or misconception, and trace backward until you hit solid prior knowledge.
3. Name the goal and the job. Decide with the student what this session fixes, and be clear whether you're repairing, helping, or prepping.
4. Rebuild from solid ground up. Start below the crack: a worked example, their wrong model made visible, then an attempt with the scaffold in place.
5. Hand over the pencil. Fade the worked example and withdraw prompts; push them to generate and explain, doing increasingly more of it.
6. Build the metacognition. Teach the self-check — "how do you know, what's next, does this answer make sense?" — so the strategy outlives the session.
7. Check independence and close the loop. End with a problem they do cold to confirm the repair held; set spaced practice, update the parent honestly, and plan the next target.

Common Tradeoffs

• Fixing the foundation vs. tonight's homework. Patch enough to unblock the deadline, but be honest the real gap is below and carve out time to go back for it.
• Telling vs. eliciting. Telling is fast and feels productive; eliciting is slow and sticks. The efficiency you feel when you explain is usually your learning, not theirs.
• Covering material vs. confirming mastery. Racing through topics recreates the classroom failure that sent them to you. Depth over coverage.
• The client's request vs. the student's need. A parent wants a higher grade by Friday; the student needs a concept rebuilt over a month. Manage the expectation honestly rather than selling a quick win that doesn't last.

Rules of Thumb

• Spend the first session mostly listening and watching them work, not teaching.
• If the explanation didn't land, the gap is lower than you think — go back a link.
• Never answer a question you can turn back into a question they can answer.
• Have them explain it back to you; if they can't say it, they don't have it.
• If you're talking more than they're working, flip it.
• Measure progress by how much help you removed, not how much you gave — and end every session with something they did entirely alone.

Failure Modes

• Re-teaching the lecture. Repeating the classroom explanation, slower and louder, without finding why it failed — same content, same crack, same result.
• Skipping the diagnosis. Teaching the visible problem and missing the prerequisite gap two layers down that caused it.
• Creating dependence. Becoming the student's external brain — they perform with you and collapse without you — and mistaking their reliance for your effectiveness.
• Over-helping / spoon-feeding. Answering before they've tried, narrating the solution, catching every error before they could catch it themselves.
• Selling the wrong job. Running homework help while the family paid for conceptual repair, or coaching test format while the student needed the foundation rebuilt.

Anti-patterns

• The narrator / lecture redux — solving the problem aloud, or re-delivering the teacher's explanation verbatim, while the student watches and nods.
• The answer machine — confirming or supplying answers the moment the student pauses.
• The crutch — making yourself permanently necessary by never fading support.
• The grade-fixer — optimizing the next test for a quick win the student can't sustain.
• The mind-reader — assuming you know the misconception instead of surfacing it from their work.

Vocabulary

• Diagnosis — the tutor's core act: identifying the specific gap or misconception from how the student works, not from the topic label.
• Prerequisite gap — a missing foundational skill that makes the current skill impossible; the thing the classroom moved past too fast.
• Misconception — a coherent but wrong mental model that must be surfaced and broken, not papered over.
• Expert blind spot / curse of knowledge — the expert's inability to see what's hard because they've automated it.
• Metacognition — thinking about one's own thinking; the self-monitoring the tutor installs.
• 2 sigma — Bloom's finding that one-to-one mastery tutoring beats class instruction by two standard deviations.
• Faded worked example — a problem progressively stripped of supplied steps until the student does it whole.
• Desirable difficulty — a struggle that, by being hard, makes learning stick.
• Think-aloud — having the student narrate their reasoning so you can see the break.

Tools

• Diagnostic questioning and the think-aloud — the primary instrument; you find the gap by making the student's reasoning audible.
• Worked examples and faded variants — the core teaching scaffold and its removal plan.
• The student's own work — graded tests, homework, the failed problem; the richest source of misconceptions.
• The school's materials — textbook, syllabus, this week's assignment, the upcoming exam — borrowed content the tutor responds to without owning.
• Concrete representations — manipulatives, diagrams, number lines, analogies — for making an abstract idea visible at the point it broke.
• Practice tests, timing, and spaced-practice plans — for test-prep the format and clock are the material; between-session retrieval makes the repair hold.

Collaboration

A tutor's relationships are smaller and more direct than a classroom teacher's, but no less load-bearing. With the student, the relationship is the precondition — trust is what lets them show you the work they're embarrassed by. With the paying client, the tutor manages expectations honestly: a grade is a lagging indicator, and repair takes longer than a deadline. With the classroom teacher, when there's contact, the tutor aligns to the methods being taught rather than introducing a competing approach. The friction lives between what the client wants (a fast score) and what the student needs (a slow rebuild); the honest tutor names that gap rather than quietly choosing one.

Ethics

A tutor is often privately hired, lightly supervised, and trusted with a child's confidence and a family's money — a position that rewards honesty and punishes its absence only slowly. The duties: tell the client the truth about what's wrong and how long it will take, even when a comforting story would sell better; refuse to do the student's work — assignment, essay, take-home exam — which is academic dishonesty dressed as help; build independence rather than the dependence that keeps the invoices coming; and protect the confidence of a child whose dignity is fragile. The gray zones — when help on graded homework crosses into cheating, when to tell a parent a goal is unrealistic, when continued tutoring is no longer in the student's interest — deserve to be faced squarely, because no one else is watching closely enough to make you.

Scenarios

The student who "doesn't get fractions." A seventh grader is failing fraction operations; the parent wants the unit fixed before the test. The novice re-teaches adding fractions — the classroom lesson again. The expert has her work problems aloud: she adds numerators and denominators straight across, and when probed can't say what a fraction is — she sees two unrelated numbers, not one quantity. The break isn't the procedure; it's two links back, in the meaning of the fraction itself. So you stop the procedure, get out a number line and paper to fold, and rebuild "three-quarters is one quantity" until it's solid; only then does the procedure go back on top, now resting on something. You tell the parent honestly the gap was conceptual and the score might lag — it did, for one test, then jumped.

The homework-help trap. A family hires you for "homework help"; every session the student arrives with that night's assignment and you spend the hour getting it done. Grades hold, then slip on tests: you've been doing triage while the underlying gaps never get touched, because the deadline always wins. You name it to the parents — the homework is getting done but the learning isn't — and renegotiate one session a week for repair, ending each with a cold, unassisted problem. Homework help and conceptual repair are different jobs, and the honest reframing bought real progress.

Related Occupations

A tutor shares the teacher's goal of causing learning but works one-to-one, diagnostically, without owning a curriculum — locating one student's specific break rather than designing instruction for a group. Teachers deliver the lesson the tutor later helps repair. Teaching assistants do related close-range support but execute a teacher's plan rather than diagnosing and choosing their own approach. Mentors do one-to-one developmental work but over a career or a craft rather than a syllabus. Students are the other half of the relationship, and the better ones the tutor is steadily training to need no tutor at all.

References

• "The 2 Sigma Problem" (1984) — Benjamin Bloom
• Mind in Society — Lev Vygotsky
• Make It Stick — Brown, Roediger & McDaniel
• How Learning Works — Ambrose, Bridges, DiPietro, Lovett & Norman
• Visible Learning — John Hattie