Differentiated Instruction in Minnesota Classrooms

Differentiated instruction begins with a common question: What is the essential learning, and what is preventing this learner or group from reaching it? It does not begin by sorting students into fixed “high,” “middle,” and “low” tracks. [S22] Minnesota's renewal statute calls for preparation in accommodating, modifying, and adapting curricula, materials, and strategies to meet individual needs and support adequate progress toward state graduation requirements. Minnesota's Standards of Effective Practice ask teachers to build on learners' assets, experiences, abilities, language, culture, family, and community values; to make instructional decisions from strengths and needs; and to plan responsive accommodations and resources. [S2; S4, Standard 1A/G and Standard 4E]

A useful planning sequence has four parts: (1) name the essential target; (2) define what success will look or sound like; (3) identify an observable barrier using more than one source of evidence; and (4) select the smallest effective change that preserves meaningful participation and progress. The change may be a universally available design feature, an accommodation, a curriculum modification, or a more individualized adaptation. The label matters less than the logic and the authority for the decision. [S26, S27]

Consider a grade-six science target: “Use evidence to explain how energy transfers in a system.” A student who understands the concept but cannot decode the textbook efficiently may face an access barrier, not a conceptual barrier. Audio-supported text, vocabulary preview, or a teacher-created diagram can preserve the same target. If the teacher instead changes the target to naming three energy words, the teacher has changed what mastery means. That may be appropriate only when it is intentionally aligned with the learner's documented goals and team decisions, not as an automatic response to slow reading. [S31]

Theoretical grounding: why a decision process rather than tracks

The research literature on differentiation, synthesized in the review by Tomlinson and colleagues, organizes learner variability along three dimensions: readiness, a learner's current proximity to one specific target; interest, the topics and modes that recruit attention and persistence; and learning profile, contextual preferences in how a learner engages, which shift across tasks and should not be treated as fixed types. Readiness is the dimension most often confused with ability, and the confusion is consequential. Ability talk implies a stable property of the student; readiness describes the distance between the student and one target on one day, a distance instruction is expected to change. Fixed tracking institutionalizes the ability reading, and the review documents both the instructional thinness and the status costs that tend to follow it. Flexible, evidence-driven adjustment operationalizes the readiness reading. [S22] The empirical base for differentiation as a total package remains uneven, in part because the practice is defined and implemented so variably across studies; the defensible claim, and the one this course makes, is narrower: specific component practices, such as barrier-matched access supports, formative evidence cycles, and explicitly taught scaffolds, each carry their own evidence, and the decision process links them coherently. [S22, S25]

A second anchor comes from the design side. The high-leverage practice framework for special education describes adapting tasks and materials as a targeted act: the teacher identifies the essential curricular element, locates the specific point of mismatch, and alters only what the mismatch requires. [S26, S27] The discipline in that account is subtraction. Every alteration the barrier does not require adds noise, muddies the evidence trail, and raises the risk of quietly lowering the target. Later lessons formalize this as the smallest-effective-change principle, and the monitoring design in Module 5 exists precisely because a small change with a clear rationale is the only kind whose effect can be read from classroom evidence. [S25, S26]

Worked example, elementary

A grade-two literacy target reads: "Retell a story, including characters, setting, and central events in sequence." Two students produce nearly identical work: fragmentary retells naming one character and one event. The behavior is the same; the barriers need not be. For the first student, a running record shows laborious decoding of the text she was asked to retell; when the same story is read aloud, her retell is complete and ordered. Her barrier is access to the story, not the retelling construct, and audio or partner reading preserves the target while her decoding receives its own instruction. [S31] For the second student, the retell remains fragmentary even after a read-aloud, but when offered picture cards of the story events he sequences them accurately and narrates from them. His barrier appears to sit in independently organizing and expressing the recalled structure, so the indicated supports are different: a taught retelling frame, rehearsal with a partner, and gradual removal of the cards. One observed behavior, two barrier analyses, two distinct and testable support decisions. Skipping the analysis and giving both students "easier books" would have served neither and would have quietly changed the target for both.

Worked example, secondary

A high-school chemistry target reads: "Use a balanced equation to model conservation of mass in a reaction." A student repeatedly submits unbalanced equations. At least three rival explanations deserve testing before any support is chosen: the student may hold the misconception that mass can appear or disappear, in which case the need is conceptual instruction with particulate models; the student may understand conservation but lose track of the multi-step symbolic bookkeeping, in which case the need is a structured workspace, a worked example placed alongside the problem, or chunked steps that reduce simultaneous processing demands; or the student may be unable to read the notation itself, in which case the need is explicit instruction in the symbol system. A two-minute diagnostic conversation, asking the student to explain what happens to the atoms, usually separates these. The support that is precisely right for one explanation is inert or counterproductive for the others, which is why this course keeps insisting that diagnosis precedes selection. [S26, S30]

Common failure modes at the diagnosis step

• Label substitution. The plan reads "he has dyslexia, so give audio" before anyone has checked what the lesson's target is. Diagnosis of the learner has replaced analysis of the task. The repair is to return to the four-part sequence: target first, then barrier, then support.
• Single-evidence inference. One low quiz score becomes a readiness verdict. Scores vary for reasons unrelated to learning, including item accessibility itself. Require a second source, such as a work sample or a conversation, before acting. [S25]
• Barrier-target conflation. The teacher treats slow reading as low comprehension, or weak handwriting as weak reasoning, and lowers the cognitive demand when only the access channel needed changing. The energy-transfer example above is this failure mode; Module 2 supplies its systematic prevention.
• Designing for the middle. No diagnosis occurs at all; one undifferentiated task is issued and the resulting struggle is attributed to students. The review literature identifies this one-size default as the pattern differentiation exists to replace. [S22]

Everything that follows in this course elaborates one branch of the sequence introduced here: Module 2 develops supports that preserve the target, Module 3 develops deliberate changes to the target, Module 4 develops individualized design, and Module 5 develops the evidence loop that tells you whether any of it worked.

Equity check. Never infer a student's learning level from disability category, home language, race, behavior, housing status, income, or a single score. Use student work, observation, conversation, progress data, and information from the student, family, and relevant team. Differentiation should expand access and challenge, not conceal lowered expectations.

PELSB's renewal rubric uses three verbs: accommodating, modifying, and adapting. The form does not define them. For this course, the terms are used as an instructional planning vocabulary, not as legal categories that override an IEP, Section 504 plan, multilingual-learner plan, district procedure, or assessment rule.

These distinctions help a teacher reason clearly, but they are not universal legal definitions. OCR states that, for Section 504 in elementary and secondary schools, the ultimate question is meaningful access and it does not require schools to distinguish “reasonable modifications” from “reasonable accommodations.” IDEA uses the terms accommodations, modifications, supplementary aids, and supports in relation to IEP-team decisions and implementation. [S8–S11, S21]

Decision boundary. A classroom teacher may offer many universally available access supports and differentiate daily instruction. The teacher must also implement required supports in a student's plan. The teacher should not unilaterally remove, substitute, or materially change required IEP/504 supports or an alternate learning goal. Questions or concerns belong in the school's documented team process. The general education teacher is an important team member who brings content knowledge, current classroom evidence, and implementation observations. [S19, S20]

The concept doing the work: construct relevance

Underneath all three terms sits a single concept from assessment theory: the construct, meaning the specific knowledge or skill a task is intended to develop or measure. A change is target-preserving when it removes construct-irrelevant demands, obstacles unrelated to what is being taught or measured, and target-altering when it removes construct-relevant ones. The National Center on Educational Outcomes' accommodations research uses exactly this distinction to evaluate testing supports, and it transfers directly to daily instruction. [S32] A read-aloud on a science item removes construct-irrelevant decoding load; a read-aloud on a decoding assessment removes the construct itself. The classification of a support therefore can never be made from the support's name. It is a three-way relation among the support, the barrier, and the construct, which is why the same support legitimately carries different labels in different lessons, and why this course keeps demanding that the target be written down before any change is chosen.

A related empirical idea sharpens the accommodation concept: the differential boost hypothesis, examined in Sireci and colleagues' review of test accommodation research. A well-matched accommodation should help students who face the barrier substantially more than it helps students who do not; a change that raises everyone's performance equally is better understood as an improvement to the task itself, and a change that only helps by removing the construct is a modification wearing an accommodation's name. The review finds the evidence pattern generally consistent with this logic for supports such as extended time, while cautioning that effects vary by student, task, and implementation, so the hypothesis is a reasoning tool and a monitoring question, not a guarantee. [S32, S37] Classroom translation: after providing a support, ask who it helped, on what, and whether the essential demand survived. Module 5 turns that question into a routine.

Worked examples of the boundary in practice

Elementary: a grade-three team wants every student to master multiplication fact fluency. For a student with a motor impairment, answering orally or on a keyboard preserves the fluency construct; the response channel was never the point. But replacing timed retrieval with an untimed matching page changes the construct, because automaticity under modest time conditions is the target. Whether that change is appropriate is a team-level instructional question, not a worksheet-level convenience. Secondary: a tenth-grade history teacher requires an evidence-based argument from documents. Providing a partially completed source-analysis organizer to a student with a documented written-language disability preserves the argumentation construct while scaffolding its executive load. Writing the thesis for the student does not. The boundary is rarely the tool; it is which cognitive work the tool leaves in the student's hands. [S26, S30]

Failure modes at the vocabulary step

• Silent modification. The most consequential error in this lesson: a target change is made informally, recorded nowhere, and reported as if the original target were met. The student's records then misrepresent readiness, and the next teacher inherits the gap without the information. If the target changed, the change must be named, authorized where required, and reflected in reporting. [S16]
• Label litigation. Colleagues spend a meeting arguing whether a support "counts as" an accommodation while no one examines the target or the evidence. OCR's own guidance signals that the labels are not the legal point in K-12; meaningful access is. Argue about the construct instead. [S11]
• Unilateral generosity. A teacher, intending kindness, waives essential requirements for a student with a plan, outside the team process. The kindness reading conceals a lowered expectation and displaces a decision that belongs to the team, including the family. [S9, S19]

An accommodation is strong when the support, barrier, and target form a visible line of reasoning. “Preferential seating” is not a complete plan. The teacher should know which barrier the seat is intended to reduce, where the student will sit, in which contexts, what the student prefers, and how the teacher will determine whether access improved. [S30, S32]

Four common access dimensions

• Presentation: captions, text-to-speech when permitted, enlarged or simplified visual layout, vocabulary preview, tactile or concrete representations, chunked directions, examples and nonexamples.
• Response: speech-to-text, keyboarding, pointing, oral response, graphic organizer, alternative communication, or a response template when the response mode is not the target.
• Setting and environment: reduced distraction, lighting/noise adjustments, accessible seating, movement access, predictable routine, and tools positioned within reach.
• Timing and scheduling: extended time, planned breaks, shorter work intervals, flexible sequence, preview/review, and scheduled check-ins.

Accommodations should not erase productive challenge. If vocabulary use is part of the target, a word bank may be appropriate while copying a complete answer is not. If decoding is the target, listening to the passage may provide content access but cannot substitute for every measure of decoding skill. A student can need different supports across activities because the targets and barriers change.

Use a five-question match test:

1. What is the essential target?
2. What evidence identifies the barrier?
3. How will the support reduce that barrier?
4. Could the support change or cue the construct being measured?
5. What indicator will show whether access and progress improved?

When an accommodation is in an IEP or 504 plan, implementation is not optional. IDEA requires the responsible teacher to be informed of the specific accommodations, modifications, and supports that must be provided. [S9] Follow local procedures for access to plans, documentation, assessment-specific rules, and concerns.

How each access dimension actually works

The four dimensions are not an arbitrary menu; each targets a different stage of the interaction between learner and task, and knowing the mechanism improves selection. Presentation supports intervene at intake: they reroute information around a blocked or overloaded channel (print to audio, spoken directions to visual sequence) or reduce the processing cost of an intact channel (chunking, vocabulary preview, decluttered layout). Response supports intervene at output: they separate the thinking from the motor, linguistic, or organizational act of expressing it, which matters whenever the expression channel is harder for the student than the thinking itself. Setting supports intervene on the environment's competing demands: noise, visual distraction, physical position, and social exposure all consume regulation capacity that would otherwise serve the task. Timing supports intervene on the schedule of demand: they redistribute the same total work so that fatigue, processing speed, or attention cycles no longer determine the quality of the evidence the student can show. [S30] The IRIS Center's accommodation framework, from which these categories derive, emphasizes a further condition that busy classrooms routinely skip: accommodations must be taught and practiced before they are needed. A text-to-speech tool introduced on assessment day measures tool fluency, not learning. [S30]

Worked examples of the match test

Elementary: a grade-four student's target is explaining ecosystem relationships. Evidence shows accurate oral explanations but written responses of one fragment, produced slowly and with visible frustration; handwriting is effortful across all subjects. The match test points to a response-dimension support: speech-to-text or oral response for the explanation task, while handwriting continues to receive its own instruction at its own time. The support does not change what mastery of the ecosystem target means, and the monitoring question is whether explanation quality now tracks his oral ability. Secondary: an eleventh grader with documented attention difficulties completes the first third of every physics problem set accurately, then deteriorates sharply. The barrier evidence suggests timing and scheduling, not physics: a planned midpoint break, or splitting the set across two sittings, redistributes demand without touching problem difficulty. If instead the teacher cut the set to the easy first third, the change would remove the multi-step problems that carry the target, a modification made by accident. The two responses look similar on paper; the construct analysis separates them. [S30, S32]

Failure modes and their diagnosis

• Support stacking. A struggling student accumulates six supports at once. When performance changes, no one knows which support mattered, and when supports conflict, no one notices. Diagnosis: the plan lists supports but no per-support barrier rationale. Repair: smallest effective change, added one at a time where feasible, each with its own indicator. [S25, S26]
• The permanent scaffold. A support installed in September is still there in May with no fading plan and no independence data. The accommodation has become the ceiling. Diagnosis: no indicator for independent performance exists. Repair: build fading criteria into the plan at adoption, as Module 5 details.
• Differential boost inversion. A support is offered to one student that would materially help most of the class (clearer directions, a worked example). This is not an accommodation problem; it is a base-design problem misfiled as an individual one, and it stigmatizes the student who receives publicly what everyone needed. Lesson 2.2 addresses the systematic answer. [S37]
• Compliance-only implementation. The plan's accommodation is technically present but functionally hollow: extended time in a chaotic room, or a scribe who edits the student's ideas. Diagnosis: access indicators do not move despite documented provision. The implementation, not the student, needs review. [S9, S29]

Universal design and individual accommodation work together. Accessible base materials reduce preventable barriers for everyone; they do not cancel a student's documented individualized supports. A captioned video benefits students who are deaf or hard of hearing, multilingual learners, students in noisy environments, and learners who want to review key terms. A student may still need an interpreter, transcript in advance, or another plan-specific support.

For digital and print materials, use the following minimum design checks:

• Use a clear heading hierarchy and meaningful reading order.
• Write descriptive link text; avoid “click here.”
• Provide text alternatives for instructional images and captions/transcripts for media.
• Do not communicate meaning by color alone; maintain readable contrast and visible focus.
• Use plain, direct directions without removing necessary disciplinary language.
• Define essential vocabulary and show at least one model or worked example.
• Make response options keyboard accessible and allow sufficient space/time.

the UDL Guidelines invite teachers to vary engagement, representation, and action/expression. Use UDL as a design prompt, not a learning-style assignment. “Visual learner” labels do not justify limiting a student to pictures. Offer multiple purposeful routes connected to the learning goal and let evidence, not preference stereotypes, guide refinement. [S13, S23]

The cognitive mechanism: why accessible design works for most learners

The design checklist above is not merely etiquette; it maps onto a mechanism described by cognitive load theory. Sweller, van Merrienboer, and Paas distinguish intrinsic load, the complexity inherent in the material relative to the learner's prior knowledge, from extraneous load, the processing imposed by the presentation itself: hunting through clutter, holding ambiguous directions in mind, integrating a diagram with a caption printed elsewhere, decoding an unfamiliar layout. Working memory is sharply limited, so every unit of extraneous load displaces capacity that could have processed the target. Accessible design is, in cognitive terms, extraneous-load reduction, which is why its benefits are broadest for learners whose working memory is already taxed by disability, by processing in a second language, or by weak prior knowledge, while costing proficient learners nothing. [S34] The same theory explains two checklist items that otherwise look arbitrary: worked examples reduce load for novices by replacing open-ended search with guided study, and chunked directions reduce the number of elements held simultaneously. [S34]

The UDL evidence base deserves an honest statement. Capp's meta-analysis of 18 studies concludes that UDL implementation improves the learning process for diverse learners; evidence tying UDL, as a whole framework, to gains on standardized achievement outcomes is thinner, and the framework's own authors describe it as design guidance rather than a packaged intervention. [S13, S23] This course therefore treats UDL as a disciplined prompt for anticipating barriers, backed by component practices with independent support, rather than as a guarantee. One boundary is firm because the evidence there is unusually clear: matching instruction to supposed learning styles has repeatedly failed rigorous tests. Pashler and colleagues' review found that studies meeting the necessary crossover design standard did not support style-matching, and the practice persists mainly on intuition. Offering multiple representations to all learners is well grounded; sorting learners into modality types and restricting them accordingly is not, and can itself become an access barrier. [S35]

Worked examples of base redesign

Elementary: a grade-one mathematics page presents twelve problems in dense rows with a decorative border, directions in italic cursive font, and color-coded operation signs. Redesign: six problems per page with white space, a plain high-contrast font, a worked example in the first position, directions in one short imperative sentence with an icon, and operation signs distinguished by symbol shape rather than color alone. Nothing about the mathematics changed; the page stopped measuring visual navigation. Secondary: a high-school biology lab handout embeds safety steps, procedure, and analysis questions in one unbroken paragraph. Redesign: numbered procedure with one action per line, safety notes set off with a consistent icon and repeated at the point of use, analysis questions moved to a separate labeled section, and key vocabulary defined in a margin gloss. Students with reading disabilities, multilingual learners, and frankly most of the class stop losing points to structure. Individual supports, such as a student's plan-required read-aloud, remain layered on top. [S13, S14, S34]

Accessible materials also protect the validity of assessment. A cluttered page, unlabeled graphic, ambiguous direction, or inaccessible response control can measure navigation or decoding burden instead of the intended knowledge. Minnesota Standard 3 asks teachers to design/adapt assessment for individual differences, minimize bias, and consider disability, communication method, culture, and primary language. [S4]

Failure modes in accessible design

• The retrofit trap. Materials are designed for an imagined typical student, and access is bolted on per student afterward, multiplying workload and singling students out. Diagnosis: the teacher maintains many parallel individual versions of the same handout. Repair: fix the base document once. [S13]
• Accessibility theater. The document passes a mechanical check (headings exist, alt text exists) but the alt text says "image," the headings are decorative, and the reading order is scrambled for a screen reader. Compliance was audited; access was not. Repair: test with the actual tools, including keyboard-only navigation. [S14]
• Simplification overshoot. In pursuit of plain language, disciplinary vocabulary and necessary complexity are stripped out, quietly lowering the target for everyone. Plain directions and rich content are compatible; the checklist item says direct language, not diluted content.

Teachers routinely vary examples, grouping, practice quantity, scaffolds, and challenge. Some changes preserve the learning target; others change what mastery means. A curriculum modification is deliberate when it names the altered target or success criteria, explains why the change is appropriate, connects to the learner's program or instructional evidence, and includes a plan for progress.

Use the target-integrity test:

• Breadth: Are fewer standards, concepts, or steps required?
• Depth: Is the student expected to explain/justify, or only identify/recall?
• Complexity: Are variables, text demands, abstraction, or independence substantially changed?
• Output: Does the response method merely provide access, or does it supply essential thinking?
• Success criteria: Is accuracy, quality, independence, or transfer judged differently?

Example: All students analyze how a character's choices influence the plot. One student uses audio text and a cause/effect organizer. The target remains analysis; those are accommodations/scaffolds. Another student identifies the character and sequences two events using picture choices. That may be a modified target. It can be instructionally appropriate when connected to the student's goals and team plan, but it must be named honestly. Calling every change an accommodation hides the learning decision and makes progress difficult to interpret.

Modification is not synonymous with “easier.” Advanced learners may require greater depth, faster pacing after demonstrated mastery, more complex sources, or open-ended transfer. Modifying for varied levels should protect intellectual dignity in every direction. The same classroom can use common themes and community membership while students work toward different documented learning goals. [S19]

The legal weight behind honest targets

The obligation to keep targets ambitious is not only an instructional value; it has a legal center of gravity. In Endrew F. v. Douglas County School District RE-1, a unanimous Supreme Court held that an IEP must be reasonably calculated to enable a child to make progress appropriate in light of the child's circumstances, explicitly rejecting a standard of merely more than trivial benefit, and stating that every child should have the chance to meet challenging objectives. [S19] The Department of Education's implementing guidance draws the classroom consequence: goals should be appropriately ambitious, and low expectations dressed as kindness do not satisfy the standard. [S20] For the classroom teacher, this converts modification analysis from a paperwork question into a professional-judgment question with stakes: a modification that reduces demand without a documented, evidence-based reason, and without a plan for progress, is exactly the pattern the Endrew F. standard exists to prevent. The same logic governs the other direction: when evidence shows a student has outgrown a modified target, leaving the modification in place is as much a target-integrity failure as imposing one casually.

Worked examples across grade bands

Elementary: the class target is adding fractions with unlike denominators. Student one receives fraction strips and a worked example alongside each problem; the target is intact and the supports are scaffolds to be faded. Student two, per an IEP goal, works on comparing unit fractions with denominators to twelve, a documented prerequisite target with its own progress measure and its own instruction; that is a modification, honestly named, connected to a plan, and reported accurately. Student three finishes early with full accuracy; the teacher substitutes fewer, harder problems requiring justification and error analysis, a depth change that preserves belonging while extending demand. Secondary: a ninth-grade English class writes a literary analysis essay. Version A supplies an annotated model and a paragraph-level planning organizer to several students: scaffolds, same target. Version B, for one student and by team decision, assesses a claim supported by two pieces of textual evidence in a structured paragraph rather than a full essay: a breadth and output modification, tied to the student's written-language goals, measured against its own criteria. Version C invites students who have already demonstrated the target to analyze an unassigned text of higher complexity with an original critical lens: an upward modification. In each case the honest question is identical: what counts as mastery here, and who decided. [S19, S22]

Failure modes in modification decisions

• Quiet curriculum shrinkage. Repeated small reductions, fewer problems this week, no essay next week, accumulate until the student is working a different curriculum no one ever decided on. Diagnosis: compare what the student was asked to do in September and in March; if the demand fell and no record explains why, shrinkage occurred. Repair: the target-integrity test on every recurring adjustment, and honest reporting. [S16]
• Modification by proxy for language. A multilingual learner receives simplified content rather than language access to grade-level content. Language proficiency was silently treated as academic capacity. The indicated supports are access supports (Module 2) plus language development, unless a team with language expertise documents otherwise. [S4, S12]
• The extension that is not one. Advanced students receive more of the same problems, or unpaid tutoring duty, labeled enrichment. Neither changes depth, complexity, or transfer, which is what an upward modification means. Diagnosis: the "extension" could be completed without any new learning. [S22]
• Grading amnesia. A modified target is taught but the report card silently scores against the original standard, in either direction. The record becomes uninterpretable for families and future teachers. Repair: align reporting with the documented target per district policy and the plan. [S16]

Poorly differentiated tasks announce a status hierarchy: one group receives rich inquiry while another fills blanks; one group uses real materials while another receives a visibly “baby” page. Strong multi-level design preserves belonging, purpose, and authentic thinking. Differences are tied to a target and evidence, not to public labels or permanent groups. [S22]

Design principles

1. Start common. Establish a shared question, phenomenon, text set, problem context, or audience.
2. Make the target visible. Students should know whether pathways share a target or use different documented goals.
3. Vary support before lowering complexity. Consider modeling, worked examples, chunking, language supports, concrete representations, collaborative rehearsal, and feedback.
4. Keep every pathway worthy. Require reasoning, communication, or authentic application, not only compliance.
5. Use flexible evidence. Regroup based on current need; do not turn a temporary scaffold into a permanent identity.
6. Plan movement. Define when support will be added, changed, faded, or intensified.

Grading and reporting must match the target actually taught and the local policy. A score should not imply grade-level mastery when a different target was assessed, nor should an accommodation reduce the grade merely because the student used it. Communicate with the appropriate team and follow the student's plan and district reporting rules. Minnesota's IEP rule requires present levels, measurable goals, progress measures, services, supplementary aids, program modifications/supports, and assessment modifications to be documented through the team process. [S16]

Student voice is evidence. Ask which supports reduce effort that is unrelated to the target, which supports feel stigmatizing, and which options increase independence. A student can decline a universally offered option; required plan supports require the school's established process rather than informal removal.

Why status effects are a design variable, not a side effect

The review literature on differentiation is blunt about the social dimension: students read grouping arrangements as information about their standing, and perceived standing feeds effort, participation, and identity over time. [S22] Two mechanisms deserve explicit management. First, expectation transmission: adults telegraph their expectations through task assignment, wait time, and the cognitive level of the questions they direct at each group, and students calibrate to what is transmitted. A pathway built only from recall tasks tells its members what is expected of them more loudly than any encouragement. Second, permanence drift: a grouping formed for one skill on one week's evidence hardens, absent an explicit movement plan, into a de facto track, reproducing the fixed-ability structure differentiation was meant to replace. The design principles above, worthy tasks in every pathway and planned movement on fresh evidence, are the countermeasures to these two mechanisms, which is why they are non-negotiable rather than stylistic. Cognitive load theory adds a caution in the other direction: for novices, high support is not condescension but efficient instruction, and the dignity problem arises only when support fails to fade as expertise grows. [S22, S34]

Worked examples of multi-level design

Elementary: a grade-three science class investigates which materials insulate best. Common launch: the same phenomenon, a melting ice cube in different wrappers, and the same driving question for everyone. Pathway one conducts the investigation with a structured data table and sentence frames for the conclusion; pathway two designs its own table and must additionally control a named variable and defend the control in writing. Both pathways answer the same question, handle real materials, and present to the same audience; movement is decided by the previous investigation's conclusion quality, not by reading group. Secondary: an algebra class works on modeling linear relationships. Common launch: one data set from the school's own recycling program. Pathway one builds the model with a partially completed table-to-graph-to-equation sequence; pathway two receives the raw data and a client-style prompt requiring a model, a prediction, and a justified statement of the model's limits. A student may complete pathway one and enter pathway two's extension in the same week; the success criteria for the shared target are posted identically for both. The design cost is real but front-loaded; the alternative cost, re-teaching a class that silently lost its lower track weeks ago, recurs forever. [S22]

Failure modes in multi-level tasks

• The visible ladder. Pathways are named, color-coded, or physically separated so that everyone knows the ranking. Diagnosis: students can tell you which group is the smart group. Repair: neutral naming, common launches, shared audiences, and public movement between pathways.
• Rich task, poor task. One pathway gets inquiry; the other gets a worksheet about inquiry. Diagnosis: apply the worthiness check to the least-supported pathway; if it requires no reasoning, communication, or application, it fails. [S22]
• Scaffold without exit. Sentence frames, calculators, or organizers appear in a pathway with no criteria for removal, so the support becomes the pathway's identity. Repair: attach a fading condition to every scaffold at design time. [S34]
• Grading by pathway. Identical work earns different grades because of the pathway it came from, or accommodation use quietly caps a grade. Both misreport learning; both violate the reporting discipline of this module. [S16]

An individualized adaptation should be traceable to a current need in a specific context. A diagnosis or eligibility category does not prescribe a single classroom method. Two students with the same category can have different strengths, communication modes, goals, sensory needs, cultural and linguistic resources, and preferences. Federal guidance emphasizes individualized decision-making; Minnesota's Standards of Effective Practice similarly require instruction that builds on strengths and needs. [S4, S9]

Build a concise learner profile with six parts:

• Strengths and interests: knowledge, language, relationships, cultural and community assets, preferred tools, persistence, curiosity.
• Current target: the specific learning or functional goal in this lesson.
• Observed barrier: what happens, in which context, and with what frequency, not an interpretation such as “unmotivated.”
• Existing plan: required accommodations, modifications, services, behavior/safety procedures, language-development supports, and assistive technology.
• Student/family/team knowledge: strategies that have worked, concerns, priorities, communication preferences, and relevant context.
• Baseline evidence: work sample, observation, brief assessment, progress measure, or student self-report.

Then use the ADAPT cycle: Anchor the target; Describe the barrier and strength; Align the smallest effective change; Plan implementation and evidence; Test, talk, and tune. The cycle discourages a “support menu” approach in which a strategy is chosen because it is familiar rather than because it matches the barrier. [S28]

Individual does not mean isolated. An adaptation can increase membership: peer-supported rehearsal with defined roles; a communication system available during group inquiry; a visual schedule that enables transition with classmates; or preteaching that prepares the student for shared discussion. The goal is access, participation, learning, and increasing agency, not adult dependence. [S24]

Calibrating intensity: the taxonomy behind the "smallest effective change"

Individualization is not a single switch but a set of adjustable dimensions, and the National Center on Intensive Intervention's Taxonomy of Intervention Intensity names seven of them: strength (the evidence behind the practice for learners like this one), dosage (frequency and duration of the support or instruction), alignment (fit between the support and the actual need, including not addressing needs the student does not have), attention to transfer (whether the design helps the skill travel beyond the supported context), comprehensiveness (how many elements of explicit instruction the support includes), behavioral or academic support (whether the complementary domain is addressed), and individualization (the presence of a data-based adjustment process). [S28] The taxonomy converts a vague instinct, "this student needs more," into a precise question: more of which dimension? A student whose support is well aligned but underdosed needs frequency, not a new strategy; a student drowning in six misaligned supports needs subtraction. The taxonomy also supplies the escalation logic this course relies on: when a well-implemented adaptation fails, the next move is a deliberate intensification along a named dimension with the data to justify it, not a lateral hop to whatever strategy is newest. [S25, S28]

Worked examples of profile-driven design

Elementary: a kindergartner's profile shows strengths in oral vocabulary and pretend play, an existing speech-language plan, and a barrier recorded as leaving the carpet during whole-group story time, most often during long stretches of teacher talk. The label "attention problem" would suggest proximity seating and reminders. The profile suggests something better aligned: shorten the uninterrupted listening intervals with embedded response opportunities (act out the verb, tell your partner the character's problem), seat the student where his frequent responses are usable, and use his play strength for retelling with props. The plan builds on documented strengths, changes the antecedent conditions rather than only the child, and defines success as participation, not stillness. Secondary: an eighth grader's profile shows strong mathematical reasoning, a 504 plan for anxiety with existing supports, and a barrier of near-total silence during required group presentations, with written work far above her spoken contributions. An aligned adaptation might sequence exposure: first a recorded solo explanation, then presenting to one chosen peer, then to a small group, with the student co-selecting each step and the plan's supports intact throughout. The target, mathematical communication, is preserved; the route is individualized; agency is a design element rather than an afterthought. Both adaptations are traceable: profile line to design decision. [S24, S28]

Failure modes in individualization

• Menu-driven support. The support was chosen because it is familiar, popular, or already photocopied, and the profile was written afterward to fit. Diagnosis: the same three supports appear in every plan the teacher writes. Repair: the ADAPT sequence run honestly, with the align step documented against the profile. [S28]
• Prosthetic dependence. The adaptation works only while an adult operates it, and independence data are never collected. The plan has purchased performance at the price of agency. Repair: write the student's role and the fading trajectory into the plan at adoption.
• Profile decay. The profile was accurate in October and is still driving decisions in April, past two review dates and a changed schedule. Diagnosis: baseline evidence is older than the current quarter. Repair: profiles carry expiration dates; Module 5's review cycle is the renewal mechanism. [S25]
• Isolation drift. Each individually sensible support (separate seating, separate materials, adult at elbow) cumulatively removes the student from peer life. Audit the plan against a membership question: during how many minutes of the day does this student learn alongside classmates? [S24]

“Provide support as needed” is not an implementable adaptation. Consistency requires operational detail while preserving professional responsiveness. A useful plan names the trigger/context, person responsible, exact action, student action or choice, materials/tools, frequency/duration, success indicator, and review date. [S29]

Implementation fidelity is not rigid sameness. It means the agreed support is present as designed and variations are documented. If a support repeatedly fails, gather evidence and use the appropriate team process. Do not quietly discontinue a required support. Conversely, do not keep an ineffective support forever just because it appears on a list; bring precise data and student/family input to review. Minnesota's current general-education teacher guide emphasizes sharing classroom evidence, participating in team decisions, and implementing documented supports rather than unilaterally revising the plan. [S15, S28]

Fidelity as a concept: what "implemented as designed" actually requires

Implementation research distinguishes several components of fidelity, and the distinctions matter in classrooms because each fails differently. Adherence asks whether the specified actions occurred; dosage asks whether they occurred at the planned frequency and duration; quality asks whether they were delivered with the intended skill and tone; and responsiveness asks whether the student actually engaged with the support. A plan can pass adherence and fail dosage (the visual sequence exists but is previewed once a week instead of daily), or pass both and fail quality (the check-in happens, delivered as surveillance rather than support), or pass all three and fail responsiveness (the student refuses a support that embarrasses him, a signal the design ignored student voice). When Module 5's data show no improvement, these four questions come before any conclusion about the student, because a support that was never really delivered has never really been tested. [S25, S29] The federal PROGRESS Center's implementation materials make the same point at the plan level: high-quality programming is a chain from clear specification through consistent delivery to informed revision, and the documentation habits this lesson teaches are what make the chain auditable. [S29]

Worked contrast: one adaptation, written twice

Consider a middle-school student whose escape-maintained shutdowns during independent writing were profiled in Lesson 4.1 terms. A vague plan says: "Allow breaks and reduce writing demands as needed." Its implementable rewrite specifies: during independent writing blocks longer than ten minutes, the student uses a two-column planning card before drafting; after each completed section he may take a self-timed three-minute break at the back table, up to two per block; the teacher checks in at minute five with one content question, not a compliance prompt; the paraprofessional records starts, break usage, and words produced on a simple tally; the student can veto the back-table location and propose an alternative; review in three weeks against baseline (two of seven blocks with any writing produced). Every noun in the rewrite is observable, every actor knows their move, the student holds real choices, and the review date converts the plan from a permanent fixture into an experiment. The rewrite took four extra minutes to draft; it is the difference between a plan and a wish. [S28, S29]

Failure modes in implementation

• Drift. The support erodes in small increments, the timer stops being set, the preview becomes "you know what to do," until the plan on paper and the plan in practice are different documents. Diagnosis: an observer following the written plan could not predict what actually happens. Repair: brief periodic self-audit against the written specification.
• Ghost supports. The plan lists supports no current adult remembers adopting, delivered by no one. These are common after staff turnover. Repair: every support names a responsible person; the review cycle retires what no one owns.
• Fidelity without sense. The opposite error: mechanical delivery of a support that evidence shows is failing, because "it's in the plan." Required supports must be implemented, but implementation and silence are not the same duty: the teacher's obligation is to deliver as designed and simultaneously to bring the failure evidence to the team. [S15]
• Documentation as performance. Elaborate logs are kept that no one reads and no decision ever consults. Documentation exists to feed the Module 5 decision cycle; if a record cannot change a decision, stop keeping it and keep one that can. [S25]

Protect privacy. Share student information only through authorized systems and with personnel who need it for their role. The reviewer artifact for this course must be de-identified: no name, student ID, birth date, school, unique circumstance, or combination that makes the learner identifiable.

A support is not successful merely because it was provided. Monitor whether it improves the intended outcome. One measure rarely tells the whole story, so use a small balanced set:

• Access: starts the task, reaches essential information, uses the communication/response mode, remains engaged with fewer avoidable barriers.
• Learning: accuracy, quality, explanation, transfer, progress toward the shared or documented alternate target.
• Independence: initiates the support, requests it appropriately, uses fewer adult prompts, or selects among effective options.
• Belonging and dignity: participates with peers, reports that the support is respectful, avoids unnecessary public separation or stigma.

Collect a baseline when feasible, identify the review interval, and use a measure sensitive enough to detect change. [S25] “Student improved” is too vague. A practical statement is: “Across four grade-seven science explanations, the student used the organizer independently in three; included claim, evidence, and reasoning in three compared with one at baseline; and reported that the organizer reduced planning load without limiting ideas.”

Interpret cautiously. Improvement after a support does not prove the support caused the change; lack of immediate improvement does not prove the student is unable. Look for patterns across tasks and talk with the student. Check whether the support was implemented as planned, the target was clear, prerequisite instruction occurred, and the measure captured the intended learning.

Minnesota Standards 3F/G ask teachers to assess individual/group performance, modify instruction, and use varied, including disaggregated, data to evaluate outcomes and adapt practice. Disaggregation can reveal access patterns, but avoid publishing small cells or identifiable learner data in course artifacts. [S4]

The evidence behind the monitoring cycle

The insistence on formal monitoring is not procedural fussiness; it is one of the better-supported practices in special education research. Fuchs and Fuchs' meta-analysis of systematic formative evaluation found that when teachers regularly measured student progress and used the data to adjust instruction, achievement rose substantially, with an average effect around seven tenths of a standard deviation, and the analysis identified two amplifiers with direct classroom translations: effects were larger when data were graphed rather than merely recorded, and larger when teachers followed explicit decision rules rather than judgment alone. [S25] The measurement tradition that grew from this work, curriculum-based measurement, adds the technical insight that useful progress measures are brief, repeatable, drawn from the actual curriculum, and sensitive to small change, properties that a unit test administered every six weeks lacks. [S36] The practical inference for a support-monitoring plan is modest and demanding at once: a one-minute counted measure, a simple chart, and a pre-stated rule ("three consecutive data points below the aimline triggers a change") will outperform an impressionistic sense of how the student is doing, however experienced the impression. [S25, S36]

Worked examples of indicator sets

Elementary: for the grade-two retelling supports from Lesson 1.1, a four-indicator set might be: access, the student initiates the retell within thirty seconds of the prompt (from a tally); learning, retell rubric score across weekly probes (charted against a baseline of two element types); independence, number of card prompts needed, aiming downward; belonging, the student participates in partner retells and reports, in a two-question check, that the cards feel like a tool rather than a mark. Secondary: for the eleventh grader's split problem sets from Lesson 2.1, the set might be: access, problems attempted beyond the midpoint; learning, accuracy on the second half compared with baseline; independence, whether the student initiates the break protocol without prompting; belonging, the student's own report on whether the arrangement is workable in the room. Neither set takes more than minutes a week to keep, and each contains at least one measure capable of showing the support failing, which is the feature that distinguishes monitoring from advocacy. [S25, S36]

Failure modes in monitoring

• Measuring provision instead of effect. The log records that the support was given, not what changed. Diagnosis: the data could be identical whether or not the student learned anything. Repair: at least one indicator must be a student outcome.
• The insensitive measure. Progress is checked with an instrument too coarse or infrequent to show change within the review window, guaranteeing "no progress" findings. Repair: brief, frequent, curriculum-sampled probes. [S36]
• Chart-free data. Numbers accumulate in a gradebook where trends are invisible. The meta-analytic finding is specific: graphing changes teacher decisions. Two minutes with a chart is part of the intervention. [S25]
• Reading noise as signal. One good Tuesday triggers premature fading; one bad Friday triggers abandonment. Decision rules exist to protect the plan from single data points, in both directions. [S25]

Differentiation is shared professional work. Minnesota Standard 7 asks teachers to communicate reciprocally with families and plan with professionals who have specialized expertise to meet unique needs. Collaboration is not a referral note that says “the student cannot do the work.” It is a focused exchange of target, strengths, evidence, supports attempted, implementation conditions, student/family perspective, and the decision requested.

Use the BRIDGE brief:

• Begin with the shared target and student strengths.
• Report the barrier in observable terms.
• Identify required plan supports and what was implemented.
• Describe response using access, learning, independence, and belonging evidence.
• Gather student/family/colleague perspective.
• End with a clear question or decision needed.

A strong brief might ask: “Should the team review the response mode because the student can explain the concept orally with the visual model but written output remains a barrier despite six weeks of implemented supports?” A weak brief asks for “an easier assignment” without naming the target or evidence.

General education teachers do not need to wait for formal eligibility before making ordinary instruction accessible and responsive. At the same time, classroom differentiation cannot replace required evaluation, special education, Section 504, English-learner services, or other mandated processes. Follow local child-find, intervention, and referral procedures when concerns persist. [S20, S21] Classroom supports also are not automatically allowed on Minnesota statewide assessments; consult the current MDE assessment resource and the appropriate team process. [S17]

Why the structure of the brief matters

The BRIDGE sequence is engineered against the predictable failure points of school collaboration. Opening with target and strengths counters the deficit frame that otherwise sets in before the second sentence and that families, reasonably, hear as a verdict on their child. Reporting the barrier observationally keeps the conversation on conditions that can change rather than traits that supposedly cannot. The implementation item, stating what the plan required and what was actually delivered, is the brief's integrity clause: a team asked to intensify support is entitled to know whether the current support was really tried, and a teacher who reports partial implementation honestly builds more credibility, not less. The evidence item imports Module 5's four-indicator discipline so that "it isn't working" arrives with data attached. The perspective item operationalizes Minnesota Standard 7's reciprocity requirement: families hold information no one else has, including what the student says at home about the support, what has worked in prior years, and what the family's own priorities are; a brief that treats the family as a recipient of conclusions rather than a source of evidence has already failed the standard. And the closing question forces the sender to know what decision is being requested, which respects everyone's time and makes the response actionable. [S4, S5]

Worked examples of collaborative escalation

Elementary: after six weeks of the retelling supports, the grade-two teacher's data show the first student thriving with audio access but the second student's retell scores flat despite verified fidelity. The BRIDGE brief to the intervention team reports the target, both strength areas, the flat trend with the chart attached, the fidelity check results, the family's observation that the student tells elaborate stories about his own day, and the specific question: does the team recommend a language screening, and should the retelling frame be intensified along the dosage dimension in the meantime? The brief does not diagnose; it routes evidence to the people authorized to decide. Secondary: the eighth grader's graduated presentation sequence from Lesson 4.1 stalls at the small-group step. The teacher's brief to the 504 coordinator and family reports the two completed steps as successes, the observable pattern at step three, the student's own account (she wants to continue but not this month), and a proposed adjustment with a review date, requesting the team's agreement rather than announcing a change to plan-adjacent practice. In both cases the teacher remains the implementer and evidence-holder while the decision sits where it belongs. [S15, S20]

Failure modes in collaboration

• The handoff brief. The communication's real message is "take this student." Diagnosis: it contains no implementation report and no question, only a description of difficulty. Repair: BRIDGE items I and E are mandatory for a reason.
• The data dump. Fifteen attachments, no synthesis, no question. Volume substitutes for analysis, and the receiving team must do the sender's thinking. Repair: one page, four indicators, one question.
• Family as formality. The family is informed after decisions rather than consulted as an evidence source before them. Beyond the relational cost, this discards the household's data and violates the reciprocity the standard requires. [S4]
• Escalation as abdication. Once the team is involved, the classroom teacher stops differentiating, waiting for specialists to solve it. Team processes add expertise and authority; they do not suspend the teacher's daily responsibility for access. [S20]

Case A: same target, access barrier. Elena, a multilingual grade-eight student, explains proportional relationships accurately with diagrams and home-language discussion. In English-only word problems she often misinterprets relational phrases, though she solves correctly after clarification. The target is mathematical reasoning, not independent English reading. Additional facts complicate the decision. Elena's unit test score was 58%, and the gradebook now flags her for intervention; item analysis shows her errors cluster entirely on multi-sentence word problems, while her computation and graph items are nearly perfect. Her English-language development plan specifies vocabulary scaffolds but says nothing about mathematics assessments. A colleague suggests moving her to the "foundations" math section, which reteaches grade-six content. Elena told her advisory teacher she likes math and finds the slower class boring. Decide which language supports preserve the target, how Elena's multilingual resources can be used as assets rather than remediation flags, what the 58% actually measured, whether the section change survives a construct analysis, and what evidence will show access improved. [S4, S12]

Case B: required individual plan. Marcus, a grade-five student with an IEP, has a documented accommodation for speech-to-text on extended written assignments and a goal addressing written organization. During a social studies project, the teacher wants every student to handwrite the final report because “writing stamina matters.” Additional facts: the project rubric allocates 80% of points to historical content and reasoning and 20% to conventions; Marcus's last two speech-to-text products were his strongest work of the year; his IEP goal targets organizational structure, which speech-to-text does not supply on its own; and the teacher's stamina rationale is not attached to any standard being assessed in this unit. Marcus has begun asking to skip drafting sessions, and his family emailed asking why he came home saying he "has to write it the hard way." Decide how the plan, the actual target, and the writing goal should guide implementation; what the rubric weighting implies about the construct; where the stamina concern could legitimately live (and in whose decision); what the teacher can decide now; what belongs in team discussion; and what response the family's email deserves this week. [S9, S19]

Case C: varied level and greater depth. Anika demonstrates mastery of the unit's grade-seven data-analysis target on the preassessment. Giving twenty more routine problems would increase quantity, not learning. Additional facts: Anika's mastery is specific, near-ceiling on center, spread, and display items, but she has never justified a choice of measure in writing, and her preassessment included no messy or real data. She is also one of three students in this position, the other two of whom dislike working together. The school has a real data set available: three years of anonymized cafeteria waste measurements collected for a sustainability grant. District policy requires that any alternate work still be graded against the same reporting standard. Design an adaptation or modification that increases depth or transfer using the taxonomy's attention-to-transfer dimension, maintains belonging without forcing an unwanted grouping, produces evidence of new learning that the reporting standard can accommodate, and states what the preassessment did not establish. [S22, S28]

For each case, use the decision record:

1. Essential target and success criteria.
2. Strengths and observed barrier/need.
3. Existing plan or rules that control the decision.
4. Selected change and its label for planning purposes.
5. Explanation of what the change preserves or alters.
6. Implementation details and student agency.
7. Access/learning/independence/belonging evidence.
8. Team/family communication or decision needed.

There may be more than one defensible answer. A high-quality response makes assumptions visible, refuses to infer from identity, and names when the information is insufficient for a unilateral decision.

The capstone is the course's professional-learning artifact. Use a real upcoming lesson, but remove all student-identifying information. The plan must be specific enough for a colleague and PELSB reviewer to see how the three criteria are taught and applied.

Required capstone sections

1. Context and target: grade/subject (or role), Minnesota/local standard, essential target, learner-friendly success criteria.
2. Asset and barrier analysis: at least two learner/group strengths and two evidence-based barriers; no diagnostic or demographic assumptions.
3. Accessible base design: at least three universal access features and a short accessibility audit.
4. Accommodation: one barrier-matched support that preserves the target; five-question match-test rationale.
5. Modification analysis: show one possible change to breadth/depth/complexity/output/success criteria; state whether it is appropriate, who authorizes it, and how it would be reported. A conclusion of “not appropriate here” is acceptable when defended.
6. Individual adaptation: de-identified ADAPT plan with precise implementation details and student agency.
7. Assessment: how students demonstrate the target; how accommodations preserve validity; how a modified target, if used, is measured accurately.
8. Monitoring: one indicator each for access, learning, independence, and belonging; baseline/review interval/decision rule.
9. Collaboration: BRIDGE brief identifying family/student/team input and any decision needed.
10. Reflection: one risk of bias or lowered expectation in the original plan and the change made to reduce it.

The capstone is complete when every section is present, the same essential target can be traced through instruction and assessment, plan-required supports are respected, and every individualized decision includes evidence and a review plan.

What distinguishes a strong capstone from a compliant one

A compliant capstone fills every section; a strong capstone coheres. Coherence shows in traceability: the barrier named in section 2 is the barrier the accommodation in section 4 addresses; the indicator in section 8 is capable of detecting that specific accommodation failing; the collaboration brief in section 9 asks a question the evidence in section 8 could actually raise. Reviewers, whether colleagues or state readers, test coherence by pulling one thread: choose the individual adaptation and follow it backward to its evidence and forward to its review plan. A strong capstone also demonstrates calibrated honesty in section 5, where the modification analysis is permitted to conclude "not appropriate here": the analysis earns that conclusion by showing what a modification would change and why the evidence does not support changing it, rather than asserting it. Finally, a strong capstone is specific about implementation: the difference between "provide sentence stems" and a stems document attached, with the fading condition stated, is the difference this course has drawn in every module between intention and design. Weak capstones fail in predictable ways: strengths sections that are compliments, barriers inferred from labels, accommodations chosen before targets were written, monitoring sections that measure provision, and reflection sections that praise the plan instead of interrogating it. The rubric weights coherence and honesty over polish, and the self-score is itself evidence of Standard 9 reflective practice. [S4, S5]