Appropriate now
- Classroom discussion and instructional planning
- Educator-reviewed evidence from this reading
- Repeated checks using carefully chosen comparable passages
- Identifying patterns that merit reteaching or another assessment
Methodology v4.1Updated August 21, 2026
Aligned by evidence, not labels.
ABZ reviews every passage as a complete evidence set: the intended standard, text complexity, word-reading demand, item reasoning, answer evidence, administration, scoring, source quality, and claim boundary. A grade label or readability number alone is never enough.
Current approved useEducator-reviewed, low-stakes classroom pilot
Not yet establishedBenchmark, diagnostic, universal screener, percentile, or state-test prediction
CURRENT EVIDENCE
What has actually been reviewed
These counts describe the current catalog—not a national norming sample or a claim that every standard has been fully measured.
45passages across K–8
184selected-response items
45constructed or oral responses
5stages in every grade
6audited controlled texts
CLAIM BOUNDARY
What “standards aligned” means here
Each item is checked against the full wording and evidence demand of one primary Common Core ELA reading standard.
Not supported yet
- Universal screening, diagnosis, retention, promotion, or placement
- National percentiles or publisher cut scores
- Mastery claims from one item or passage
- Automatic alignment to every state or national school system
Jurisdiction rule: Common Core is the current base framework. TEKS, Florida B.E.S.T., provincial frameworks, and national curricula require a separate, versioned crosswalk reviewed by someone familiar with that system. A partial match must be labeled partial.
PASSAGES & QUESTIONS
Seven evidence checks before assignment
Alignment is reviewed at the item level. A passage can mention the right topic and still fail if the question does not require the named construct.
Quantitative + qualitative + reader/task
We record Flesch–Kincaid, ARI, word and sentence measures, then review meaning, structure, language, knowledge demands, supports, and intended administration.
The full standard—not just its code
The prompt, passage, options, key, evidence, and response rubric must require the cited skill, including boundaries such as text type or “two or more texts.”
DOK belongs to the whole task
Selected-response items normally use DOK 1–2. K–2 direct retells and explanations remain DOK 2. Grades 3–8 reach DOK 3 only through strategic integration, analysis, application, or evaluation.
One defensible key with located evidence
Every option is unique. The key has a rationale and a sentence, line, paragraph, image, or whole-text locator. Trick distractors and ambiguous keys fail review.
Visible educator scoring criteria
K–2 uses a 0–3 response rubric; grades 3–8 use 0–4. Rubrics are internally developed and remain pending an external inter-rater study.
Traceable and accessible
Informational texts retain authoritative sources and a factual-review date. Images require provenance, accuracy, meaningful alt text, and review whenever they supply assessed evidence.
Grade floor plus bounded length adjustment
Grade sets the minimum type size and maximum line measure. Very short passages receive larger print; sustained passages never shrink below the grade floor. Stage changes complexity—not print size.
18 / 45
passages currently trigger ABZ's formula-review flag because FK or ARI is at least two grade levels above placement. A flag is not an automatic failure—domain vocabulary can inflate formulas—but it requires documented human placement and student field testing, especially at entry stages.
PASSAGE DISPLAY
Readable by grade, responsive to length
The same versioned profile appears in the educator preview, student read-aloud screen, and question reference. Typography supports access; it does not determine placement.
Grade is the floor
Kindergarten begins at 2.00rem with a 24-character wide-screen measure. Type size and spacing step down gradually through Grade 8, where the floor is 1.00rem and the measure is 72 characters.
Short text receives more space
Passages of 35 words or fewer add 0.25rem; passages of 36–75 words add 0.12rem. Longer text keeps the grade base rather than being compressed to fit.
Consistency rule: Stage A–E and readability formulas never reduce print size. Paragraphs remain visible, browser zoom remains available, mobile screens wrap sooner, and any future profile change receives a new display version.
WITHIN-GRADE PATHWAY
Grade destination plus Stage A–E
Stages describe passage and task complexity—not a child, month, percentile, Lexile level, or conversion to another publisher's system.
| Stage | Educator label | Purpose |
|---|---|---|
| AEntry | Entry / prerequisite | Least complex grade-pathway work with substantial support or tight code control where applicable. |
| BBeginning | Beginning of grade | Accessible language, explicit relationships, and structured support. |
| CDeveloping | Developing within grade | Moderate support while building toward the full grade expectation. |
| DLate grade | Late in grade | Most grade-level demands with limited support and stronger evidence integration. |
| EBridge | End-of-grade / bridge | Strongest passage-level work in the pathway and preparation for the next grade. |
Placement uses word recognition, language, meaning, structure, knowledge, and task together. Adjacent stages intentionally overlap in length because a shorter text can be conceptually or linguistically harder than a longer one. Five passages per grade provide a usable pathway, not deep genre or standards coverage.
FOUNDATIONAL READING
Word reading and comprehension stay separate
ABZ does not average two different constructs into one opaque “reading level.”
Foundational word reading
The student reads a phonics-controlled text. Each running token is classified as a cumulative pattern, permitted irregular word, pre-taught story word, or disclosed out-of-sequence word.
Language comprehension
The educator may read or support a richer text. Results describe understanding, not independent decoding mastery. Controlled-text questions are delivered orally so question decoding does not contaminate the passage result.
Current controlled-text range90–100% instructionally controlled
ABZ code steps are an internal instructional scope, not an official Orton-Gillingham sequence. Approved wording: Designed around structured, explicit, sequential, cumulative literacy principles; not Orton-Gillingham certified.
ORAL READING RATE
Educator-reviewed whole-passage WCPM
The calculation is accurate for the recorded ABZ passage when the educator completes the running record. It is not automatic speech scoring.
correct words = source words attempted − scored errorswhole-passage WCPM = correct words ÷ exact elapsed milliseconds × 60,000accuracy = correct words ÷ source words attempted × 100- The timer begins only when the student presses Start reading; microphone setup is excluded.
- Reading is continuous with no pause control. The educator marks the last source word attempted, and later words are excluded.
- The passage counter and scorer use the same tokenizer, including em-dash-separated words.
Scored: omissionsubstitutionmispronunciationword supplied after 3+ secondsCorrect: self-correction within 3 secondsNote only: insertionNote only: repetition
Interpretation boundary: This is a passage-specific ABZ rate, not a DIBELS score, percentile, diagnostic result, or national benchmark. Raw WCPM changes across unlike forms may reflect passage difficulty as well as student performance.
RELEASE & VALIDATION
What is complete—and what comes next
Product leadership cannot assign “validated” by declaration. Each status requires documented evidence for a named use.
Internal alignment QA complete
Catalog structure, standard tags, keys, rationales, source logs, word counts, code audits, administration rules, and automated checks are complete for the pilot.
Independent educator review
Qualified reviewers who were not the original writers must review grade placement, content, alignment, accessibility, fairness, and response rubrics.
Student field testing
Cognitive interviews and field data must examine interpretation, difficulty, discrimination, distractors, reliability, subgroup behavior, and passage-form effects.
Comparability and validation
Multiple comparable forms, scorer agreement, and a defensible linking plan are required before benchmark, growth, norm, diagnostic, or predictive claims.
PRIMARY FRAMEWORKS
Sources educators can inspect
ABZ stages and length targets are internal design tools. These sources inform the method but do not certify the product.
Related ABZ records: Review the public Content & Standards Registry for the current working CCSS map and Evidence & Research for ABZ's evidence boundaries and next validation steps.