Assessment, Instruction, and the Learning Environment
Domain 2 treats assessment, instructional design, and classroom climate as one system: what you want students to know determines how you teach it and how you find out whether they learned it.
Assessment rules
- Purpose first: diagnostic (before), formative (during, feedback-driven), summative (after, judgment).
- Validity means the inferences drawn are appropriate; reliability means consistency. A test can be reliable and still invalid. Bias is any construct-irrelevant barrier - language load, cultural reference, format.
- Build a table of specifications aligned to Ohio's Learning Standards; sample content and cognitive demand deliberately.
- Item quality: one idea per item, plausible distractors, no "all of the above," unambiguous stems.
- Rubrics: analytic for criterion-by-criterion feedback, holistic for overall judgment; share criteria before the task.
- Grades should reflect achievement of standards, not compliance or effort.
Instruction
Backward design sequences desired results → acceptable evidence → learning plan. Gradual release moves from modeled instruction through guided and collaborative practice to independent application. Cooperative learning requires positive interdependence plus individual accountability. Differentiation adjusts content, process, or product; Universal Design for Learning supplies multiple means of representation, action and expression, and engagement.
Learning environment
Procedures are taught and rehearsed like content. Proximity, withitness, and the least-intrusive intervention preserve instructional momentum. Positive behavioral interventions and supports (PBIS) organize universal, targeted, and intensive tiers.
Worked example
A chair asks a teacher to "cover" linear functions in three weeks. Reasoning: start from the standard and its cognitive demand; write the summative blueprint first; add two formative checkpoints (a four-item exit ticket after modeling, a performance task at week two); plan gradual release and a reteach day keyed to exit-ticket errors. Evidence drives the calendar, not page counts.
Traps
- Swapping reliability and validity definitions.
- Reteaching the whole unit without item-level analysis.
- Calling group work cooperative when no individual accountability exists.
- Escalating discipline publicly instead of using proximity or private redirection.