Assessment and Evaluation
What to know
Assessment questions test whether you can choose the right instrument and read scores correctly. Distinguish norm-referenced tests (compare a student to peers; yield standard scores, percentiles, stanines, age/grade equivalents) from criterion-referenced tests (mastery of specific skills). Reliability is consistency (test-retest, alternate form, internal consistency, interrater); validity is accuracy for the intended purpose (content, construct, criterion/predictive). Standard scores have a mean of 100 and SD of 15, so 85 is one SD below the mean, roughly the 16th percentile - never "85% correct." The standard error of measurement creates a confidence band, so score differences smaller than the SEM are not meaningful. Know purpose-based types: screening (brief, flags risk), diagnostic (pinpoints deficits), progress monitoring (curriculum-based measurement, frequent, sensitive to short-term growth), and outcome/accountability.
Traps: age and grade equivalents do not mean the student performs at that grade level; percentile is not percent correct. Severe discrepancy is no longer required for identification - RTI/MTSS data and patterns of strengths and needs drive eligibility. Nondiscriminatory assessment means testing in the child's native language, using multiple measures, and ruling out lack of instruction or limited English proficiency as the cause of low achievement. Typical timelines: initial evaluation and eligibility determination within 60 school days of parental consent, reevaluation every three years or when conditions warrant, and informed written consent before initial evaluation.
Rule of thumb: match the purpose word in the stem to the tool - "predict" or "compare to peers" means norm-referenced; "did the student master this" means criterion-referenced; "is the intervention working week to week" means CBM progress monitoring.