The Tutor Handbook · Volume 0188 · Series ID THB-0188
The Tutor Handbook: Complete Series Index
Two examples can make a distinction visible—or hide it under twice the detail
A tutor puts two worked solutions side by side. The learner is asked to compare them.
One uses factorisation. The other uses the quadratic formula. Both reach the same answer. The tutor expects the learner to notice when each method is efficient. Instead, the learner notices that one solution is shorter, one contains a square root, and one has more writing. The comparison has produced observations, but not the distinction that matters.
In another lesson, a tutor compares two English paragraphs. One is well adapted to audience; the other is formally correct but tonally wrong. The tutor highlights half the page, asks six questions and explains the answer before the learner has made a private judgement. The learner agrees with every point. On a third paragraph, the same learner cannot identify the tonal mismatch without help.
Comparison is powerful because differences become easier to see against a carefully chosen neighbour. It is also demanding because two cases double the amount of surface information the learner can attend to.
The Contrast-Case Gate asks a practical tutoring question: which two examples should be placed side by side, when is the learner ready to compare them, how much guidance should the tutor provide, and what later evidence shows that the learner noticed the decisive difference rather than merely following the tutor’s comparison?
The job is not “use more examples”. It is to choose a pair whose relationship does useful instructional work.
Quick answer
Use a contrast pair when the learner needs to discriminate between two confusable methods, categories, explanations, representations or answer qualities and already has enough prior knowledge to inspect the cases meaningfully.
Choose the pair so that one or two decisive features change while irrelevant features are controlled as far as practical. Present the cases close enough together that the learner can compare them directly. Give the learner a private first look before the tutor announces the key difference. Ask focused comparison questions rather than a vague “What is different?” when the search space is large. Align notation, layout or terminology when misalignment would create unnecessary load.
Then ask the learner to state the distinction in a form that can travel: a decision rule, causal boundary, quality criterion or structural relation. Finally, present a third fresh case. If the learner can classify, choose or explain the third case without the original pair carrying the answer, the comparison has a better receipt.
Do not compare simply because two examples are available. The pair should make a decision easier to learn and harder to fake.
This article owns pair selection, not analogy in general
The broader cognitive mechanism of mapping structure across different-looking problems already has an owner in the MindOS learning manual Analogical Mapping State | Two Problems Can Look Different and Still Have the Same Skeleton. That learner-side manual explains how structural relationships can be mapped across cases.
The Tutor Handbook already has an Example-Variation Gate, which asks what should vary and what should remain invariant across examples. It has an Erroneous-Example Gate, which handles deliberately wrong solutions. It has an Interleaving Readiness Gate and an earlier Mixed Set for method selection across practice.
The Contrast-Case Gate is narrower.
It owns the tutor’s decision to put this case beside that case now because their relationship exposes a discriminating feature the learner needs.
That pair can contain two correct strategies, two confusable categories, two answers of different quality, two representations of the same structure, or—when readiness permits—one correct and one incorrect example. The instructional question is the pair and the comparison job.
Why comparison is more than “spot the difference”
A useful contrast changes what the learner is able to notice.
If a learner sees only examples of direct proportion, they may learn the surface pattern of those examples without learning the boundary between direct and inverse relationships. If they later see a carefully matched inverse case beside a direct one, the decision variable can become more visible.
If an English learner sees only strong situational-writing openings, they may imitate phrasing without understanding why the language suits the audience. A matched opening that is grammatically correct but too informal can expose the audience decision.
If a Science learner repeatedly sees experiments with one changed variable, they may memorise “change one thing” without understanding why two superficially similar investigations answer different questions. A contrast can reveal which variable is manipulated, which is measured and how the claim changes.
The power comes from relational information: not simply what Case A contains, but what becomes meaningful because Case B is next to it.
Research makes the what–when–how problem explicit
Bethany Rittle-Johnson, Jon Star and Kelley Durkin’s 2020 review, How Can Cognitive Science Research Help Improve Education? The Case of Comparing Multiple Strategies to Improve Mathematics Learning and Teaching, synthesises research on comparison in Mathematics and emphasises that learning depends on what is compared, when the comparison occurs and how learners are supported to compare.
The review discusses benefits from comparing multiple strategies for the same problem and comparing confusable problem types. It also makes clear that comparison is cognitively demanding. Learners need enough knowledge and support to identify meaningful similarities and differences rather than drowning in surface features.
The authors discuss design features such as side-by-side presentation, aligned terminology or visual cues, focused explanation prompts and subsequent summarisation. They also describe implementation work in classrooms where simply introducing comparison did not automatically produce strong use; curriculum integration and professional support mattered.
This is a review and translational research discussion, not a direct trial of this tutoring gate. It gives a strong reason to treat comparison as a designed instructional event rather than a decorative activity.
Start with the decision the learner must learn
Before choosing the two cases, write the decision in one sentence.
“Learner must distinguish proportional from additive change.”
“Learner must decide whether a sentence reports evidence or makes an inference.”
“Learner must see why these two algebra methods are equivalent but one is more efficient under this structure.”
“Learner must distinguish a relevant supporting detail from a detail that is merely related to the topic.”
Without that sentence, tutors often choose examples because they are convenient, visually interesting or already available. The comparison then has too many possible lessons.
A strong pair is designed backwards from the discrimination.
Change the decisive feature, control the distractions
Suppose the tutor wants a learner to distinguish mean from median as a better summary under skew. A weak pair uses two completely different data sets: one about salaries with ten numbers and one about test scores with twenty numbers, different scales, different graph types and different questions. The learner has too many differences to inspect.
A stronger pair can begin with nearly the same data. In one set, the values are clustered. In the second, one extreme value changes. The tutor asks what happens to mean and median and which statistic remains representative of the typical case.
The changed feature carries the lesson.
The same principle applies beyond Mathematics. If the tutor wants an English learner to notice tone, keep purpose and content similar while changing audience-sensitive language. If the tutor wants a Science learner to distinguish observation from inference, use two statements about the same phenomenon so the key difference is epistemic rather than topical.
Perfect control is not always possible. The aim is to reduce irrelevant differences enough that the important one has a chance to become visible.
The learner needs enough prior knowledge to compare
Comparison can arrive too early.
If neither case is understood, the learner is not comparing. They are trying to decode two unfamiliar objects at once.
A novice who has never understood factorisation and has never seen the quadratic formula is unlikely to learn method choice by comparing full worked solutions immediately. The tutor may first need to establish each method separately enough that the learner can recognise its steps and purpose.
Likewise, comparing two sophisticated essays can overload a learner who cannot yet identify the claim in either one. Comparing two circuit explanations is not useful if basic current and voltage relationships are missing.
The gate therefore asks a readiness question: does the learner know enough about the components to learn from the relationship between them?
If not, teach the parts first.
This protects comparison from becoming “discovery” theatre where the tutor knows the distinction and the learner is expected to infer it from noise.
Side by side usually beats memory-to-memory comparison
Working memory matters. If the learner reads Case A, turns the page and later reads Case B, much of the comparison depends on remembering the first case accurately.
Placing cases side by side—or otherwise keeping both accessible—reduces that unnecessary memory burden. The learner can inspect the actual details rather than compare one visible example with a vague memory.
This does not mean every contrast must be physically adjacent forever. Once the distinction is learned, delayed comparison can be useful because memory itself may be part of the target. But during instruction, when the job is to notice a relationship, simultaneous access often makes the comparison cleaner.
For online tuition, this can mean split-screen presentation or a shared document. For paper, it may mean a two-column page. For oral discussion, the tutor can write the crucial features of both cases before asking for a judgement.
The layout should serve the relation.
Align notation when notation is not the lesson
Two equivalent Mathematics methods can look unrelated because one uses different variable names, one writes fractions vertically and the other inline, or the order of steps differs for irrelevant reasons.
Two English examples can use different topics, lengths and vocabulary, making tone harder to isolate.
Two Science explanations can use different labels for the same process, adding language noise to a conceptual comparison.
When those differences are not the learning target, align them.
Use common labels, comparable step positions and visual correspondence where sensible. The aim is not to make the cases artificially identical. It is to stop the learner spending attention on a formatting difference that the tutor does not intend to teach.
Later, once the structure is understood, the tutor can vary notation deliberately to test transfer.
Private noticing before public comparison
In a three-learner tutorial, the fastest speaker can destroy the comparison for everybody else.
The tutor shows two solutions and asks, “What is the important difference?” Beatrice immediately says, “This one factorises because the coefficient pattern is easy.” Alicia and Ciara now know what to look for. Their later agreement is no longer independent evidence of noticing.
A small protection is private first response.
Ask every learner to mark one important similarity and one important difference before discussion. Or ask each to choose which method they would use and write one reason. The private response can take thirty seconds.
Then open the comparison.
This preserves the social value of discussion while keeping an initial sample of each learner’s perception.
It also helps the tutor decide whose contribution will add information rather than simply reward speed.
Vague comparison questions create vague answers
“What is similar and different?” is sometimes useful. It can also produce a list of surface features.
One solution has four lines. The other has five. One paragraph says “however”. The other does not. One graph is blue. The other is red.
When the search space is large, use a focused prompt.
“What decision changes between these two methods?”
“Which sentence makes the audience relationship different?”
“What evidence is available in both cases, and what conclusion changes?”
“Where does the first irreversible error appear?”
“What stays invariant when the representation changes?”
Focused prompts do not necessarily give away the answer. They define the dimension of comparison so the learner can do meaningful work inside it.
The right level of focus depends on prior knowledge. More expert learners can handle broader comparison questions because they have better internal categories for what matters.
Composite case: two correct Mathematics methods
The following case is fictional and constructed for teaching.
Alicia can solve simultaneous equations by elimination and substitution. She performs both accurately when the method is named. On mixed questions, she chooses slowly and sometimes selects a cumbersome method.
Her tutor could assign twenty mixed problems. Instead, he selects two matched systems.
In Case A, one variable already has coefficient 1 and is isolated easily. In Case B, coefficients align for elimination after a simple multiplication. Both cases are presented side by side using the same variable names and layout.
Before discussion, Alicia marks which method she would choose for each and why. She selects substitution for both because “I understand it better”. The tutor does not tell her she is wrong. He asks her to estimate the number and complexity of transformation steps each choice would require.
Alicia works through the comparison and notices that method validity and method efficiency are separate questions. She then writes a provisional decision rule: substitution is attractive when a variable is already easy to isolate; elimination is attractive when coefficients are already aligned or can be aligned simply.
The tutor then presents a third fresh system that does not exactly match either original. Alicia must choose without the pair visible.
The lesson did not prove a universal rule. It built a more discriminating method-selection cue.
Compare two strong cases when the learner needs flexibility
Contrast does not require one good and one bad example.
Two correct strategies can reveal that a problem has multiple legitimate routes. This is useful when learners have become over-dependent on one method or assume the model answer is the only valid answer.
Two strong English introductions can reveal different ways to establish purpose. Two correct Science explanations can use different evidence while preserving the same causal model. Two Mathematics solutions can expose a trade-off between efficiency and transparency.
Comparing two strong cases can prevent the learner from turning instruction into imitation.
The tutor’s job is to choose cases where the difference teaches something transferable rather than merely showing personal preference.
Correct versus incorrect comparison needs readiness
Wrong examples can be powerful because they force discrimination. They can also seed or strengthen a misconception.
The Erroneous-Example Gate owns the detailed decision about when deliberately incorrect solutions are safe and useful.
For contrast-case design, the rule is simple: do not assume that putting a correct and incorrect answer side by side automatically clarifies the difference. The learner needs enough correct knowledge to inspect the error and enough structure in the task to locate why it fails.
If the learner is still building the basic model, compare two correct examples first or provide a correct anchor before introducing the error.
When the wrong case is used, ask for the failure mechanism—not merely “Which one is wrong?”
Composite case: an English audience contrast
This case is fictional.
Ciara writes grammatically accurate emails but often uses the same tone for teachers, close friends and formal organisations.
Her tutor chooses two short openings for the same request. Both communicate the same factual content. One begins with language appropriate for a school administrator. The other sounds like a message to a friend. Vocabulary difficulty and sentence length are kept similar.
Ciara first labels each opening’s likely audience. She identifies both correctly but says the difference is simply that one is “more formal”.
The tutor asks a narrower question: which words change the relationship between writer and reader? Ciara highlights greeting, modal choice, request phrasing and sign-off. She then rewrites a third opening for a different audience while preserving the purpose.
The contrast has moved from a vague category—formal versus informal—to a set of audience-sensitive language decisions.
A later writing task tests whether she can apply those decisions without the pair beside her.
Comparison can diagnose as well as teach
A well-chosen pair can reveal what the learner is using as a decision cue.
Suppose two Science questions look similar but only one supports a causal claim. Ask the learner which permits the stronger conclusion and why. Their explanation can reveal whether they are attending to variable control, sample design, correlation language or simply familiar wording.
Suppose two Mathematics problems require the same underlying structure but have different contexts. Ask whether they belong to the same method family. A surface-driven learner may separate them because the stories differ.
Suppose two English sentences are both grammatically correct but only one answers the question directly. Ask which better meets the task purpose. A learner focused on sentence polish may ignore relevance.
The pair acts as a discriminating question.
This is especially useful for Class 3 Diagnostic Tutor work, where the aim is not merely to teach more but to discover which feature is currently controlling the learner’s decision.
Composite case: Science observation and inference
This case is fictional.
Denise knows the definitions of observation and inference. She can recite them accurately. In actual Science questions, she still mixes them.
The tutor presents two statements about the same setup:
“Water droplets are visible on the outside of the cold container.”
“The water droplets came from water vapour in the surrounding air.”
Denise says both are observations because both are true.
The tutor asks what could be recorded directly with the senses or an instrument and what requires an explanatory model. Denise now separates direct evidence from explanation.
A third statement is added: “The mass measured on the outside surface increased by 0.6 g.” Denise classifies it as observation despite the numerical form. A fourth statement introduces a mechanism. She classifies it as inference.
The comparison did not teach the vocabulary again. It changed the cue Denise was using to classify statements.
The third case is the receipt
Comparison can create immediate recognition. The original pair itself contains strong cues.
If the learner only succeeds while the two examples remain side by side, the tutor does not yet know whether the distinction has become independently usable.
A third case changes that.
The third case should be fresh enough that the learner cannot simply copy a surface rule, but close enough that the same decision is genuinely relevant. The learner classifies, chooses or explains it without being told which original case it resembles.
A delayed third case is stronger when durability matters. A changed-context case is stronger when transfer matters.
The third case is not always required in the same lesson. If the pair is being used during initial teaching, supported comparison may be the appropriate endpoint. But before making a stronger claim about independent discrimination, the tutor needs evidence beyond the original pair.
Contrast and interleaving solve different problems
Contrast puts two cases into a deliberate relationship so the learner inspects a difference or common structure.
Interleaving mixes problem types across practice so the learner has to select a method without a topic label. The Interleaving Readiness Gate asks when that selection demand is appropriate.
A tutor may use contrast first to make a boundary visible, then later use interleaved practice to see whether the learner can recognise the boundary without the cases being paired explicitly.
This sequence is often more interpretable than throwing confusable problems into a mixed set before the learner knows what distinction to inspect.
Comparison can teach the discriminating feature. Interleaving can later test whether that feature controls selection.
Contrast and example variation solve different problems
The Example-Variation Gate asks what should vary across a family of examples so the learner can abstract the invariant without the target changing into another skill.
Contrast-case design uses two examples more deliberately. The pair is chosen because their relationship exposes a decision.
A sequence can use both. Several examples may vary surface context while preserving the same deep structure. Later, a near-neighbour case with a different deep structure can be placed beside one of them to clarify the category boundary.
The tutor should know which instructional job is active. Otherwise “varied examples” can become an undifferentiated pile.
Three learners create both an advantage and a hazard
Small-group tuition is unusually well suited to comparison because learners can contribute different observations. One notices a structural similarity, another spots a boundary, and a third challenges the proposed rule.
The hazard is answer leakage.
If the quickest learner announces the decisive difference immediately, the others lose the chance to notice it independently. If the tutor then asks, “Do you agree?”, group fluency can masquerade as individual understanding.
Use a short private phase. Then distribute roles. One learner can propose the difference. A second can test it against both cases. A third can produce a counterexample or identify a limit.
After discussion, each learner should face a small individual decision when the target requires independent discrimination.
The group makes comparison richer. It should not make the evidence less honest.
Do not compare too many dimensions at once
A tutor can become enthusiastic about comparison and build a table with twelve rows: method, speed, accuracy, notation, cognitive load, elegance, transfer, common errors, examiner expectations, representation, prerequisites and checking.
The learner fills the table and remembers none of the decision.
Comparison has a bandwidth limit.
Choose the dimensions that matter for the present learning job. A novice may need one decisive contrast. A more advanced learner may compare efficiency, generality and failure conditions simultaneously.
The tutor can return later for a second layer.
Depth beats decorative completeness.
Do not make the tutor’s preferred method the hidden answer
When comparing two valid methods, tutors can unconsciously design the pair so their preferred method wins every time.
The “alternative” is shown in a cumbersome example. The preferred method is demonstrated with cleaner numbers. The tutor asks questions that praise efficiency but never discusses transparency or generality.
The learner concludes that comparison means discovering the tutor’s favourite route.
A fair comparison declares the criterion. Efficient for what? Transparent for whom? Robust under which values? Compatible with which school method? Easy to verify under examination conditions?
The Alternative Method Gate protects legitimate learner methods. Contrast-case teaching should support that same professional discipline.
Parents can use comparison without turning home study into interrogation
At home, parents often ask, “What is the difference between these two?” That can be useful when the child already has enough knowledge and the comparison is focused.
The parent should resist explaining the distinction immediately. Give the learner a moment to mark or say one important difference. If the learner cannot, the problem may be missing knowledge rather than insufficient effort.
Parents should also avoid manufacturing trick comparisons merely to catch the child. The purpose is to make a meaningful boundary visible, not to reward vigilance against hidden traps.
In most cases, the tutor remains the better owner of pair selection because the pair should connect to the learning route and later evidence.
Failure modes
The decorative-pair failure. Two examples are shown together because comparison sounds sophisticated, but the pair has no clear decision job.
The many-differences failure. Too many irrelevant features change at once, so the learner cannot tell which difference matters.
The no-prior-knowledge failure. Learners are asked to compare two unfamiliar cases and end up decoding rather than relating.
The answer-leakage failure. The tutor or fastest peer names the decisive feature before others have a chance to inspect it.
The vague-prompt failure. “What is different?” produces surface observations when a focused comparison dimension was needed.
The false-equivalence failure. Two methods or answers are presented as equally valid when one violates the task, evidence or curriculum requirement.
The favourite-method failure. The tutor designs the comparison to prove a personal preference rather than reveal a genuine decision boundary.
The table-overload failure. Learners complete an elaborate comparison organiser that consumes attention without sharpening the target distinction.
The no-third-case failure. Success with the original pair is treated as proof that the learner can discriminate independently later.
The confusion-seeding failure. An incorrect example is introduced before the learner has a sufficiently stable correct model.
A practical contrast-case design sequence
Begin with the learner decision: what must they discriminate, choose or explain?
Choose two cases whose relationship makes that decision visible. Reduce irrelevant differences. Check that the learner has enough prior knowledge to understand both cases. Present them close together. Give private noticing time. Ask a focused comparison question. Let learners state the distinction. Refine the language only as much as necessary.
Then ask for a generalisation that is modest enough to be true.
Finally, offer a fresh third case. If the learner’s decision holds, the comparison has begun to travel. If it fails, inspect the rule they formed. The learner may have attended to a surface cue rather than the intended structure.
That failure is useful evidence. It tells the tutor what the pair taught in practice, not merely what it was designed to teach.
Evidence boundaries and sources
Rittle-Johnson, Star and Durkin’s 2020 review, How Can Cognitive Science Research Help Improve Education? The Case of Comparing Multiple Strategies to Improve Mathematics Learning and Teaching, was published online on 12 November 2020 in Current Directions in Psychological Science. It reviews research on comparing strategies and problem types and emphasises that effectiveness depends on what is compared, when comparison occurs and how learners are supported. Much of the evidence discussed concerns Mathematics; transferring the design principles to English and Science in this article is a reasoned instructional application, not direct experimental proof.
The general mechanism of learner analogical mapping remains owned by the existing MindOS manual. The examples here use comparison as a tutor-side selection and orchestration problem.
No claim is made that every useful contrast must be simultaneous, that every comparison needs a third case immediately, or that one pair can establish a durable concept. Those are professional design rules whose value depends on target, learner knowledge and later evidence.
The end state
A contrast pair should leave the learner with a better decision, not merely a better memory of two examples.
The learner should become more able to say: these cases look similar, but this feature changes the method; these answers are both fluent, but this one fits the audience; these observations are both true, but only one supports that conclusion; these strategies both work, but their advantages change with the structure.
That kind of discrimination is one of the foundations of transfer.
The tutor earns it by choosing the pair carefully, keeping the comparison cognitively manageable, protecting the learner’s first noticing, and checking a fresh case afterwards.
That is the Contrast-Case Gate.
Do not ask two examples to teach everything. Put them side by side because there is one important relationship the learner is now ready to see.
