Reading Assessment and Instructional Decision Making
This domain asks whether you can select, administer, and interpret literacy assessments and then match instruction to the data. Expect items that give you a student profile and ask what the numbers mean or what to do next.
Four assessment purposes
- Screening: brief, universal, identifies students at risk (fall/winter/spring benchmarks).
- Diagnostic: in-depth, for flagged students; pinpoints specific gaps via phonics inventories, error analysis, or an informal reading inventory.
- Progress monitoring: short, frequent, equivalent forms, sensitive to small growth (curriculum-based measurement, oral reading fluency).
- Outcome: annual accountability measures.
Reading the numbers
- Accuracy = words read correctly ÷ words attempted.
- Informal reading inventory levels: independent 95–100%, instructional 90–94%, frustration below 90%.
- Fluency has three parts - accuracy, rate (words correct per minute), and prosody. WCPM benchmarks differ by grade and season.
- In miscue analysis, a self-correction that resolves an error is not counted as an error; omissions, substitutions, and insertions are.
Matching instruction
- Word-level: explicit, systematic, cumulative, multisensory routines.
- Comprehension: text-structure instruction, question-answer relationships, reciprocal teaching, think-alouds, and strategy modeling.
- Vocabulary: Tier 1 (everyday), Tier 2 (high-frequency academic words worth teaching directly), Tier 3 (domain-specific).
Worked example
A second grader reads 42 WCPM at 89% accuracy, word by word, with flat expression and no self-corrections; the winter benchmark is 50 WCPM at 95%. Reasoning: the student is below benchmark on both rate and accuracy, and prosody is weak, so this is a fluency and decoding profile rather than a comprehension profile. The appropriate move is repeated reading with a model, phrase-cued reading, and targeted phonics review - not a comprehension-strategy unit.
Traps that cost marks
Treating a screener as a diagnosis; drawing conclusions from one data point; counting self-corrections as errors; responding to slow reading with comprehension instruction when accuracy is also low.