Assessment and Data-Based Decision Making
Types and technical quality
Assessment serves screening, diagnosis, progress monitoring, and summative decisions. Norm-referenced tests compare students to a sample; criterion-referenced tests compare to a standard. Reliability is consistency; validity is evidence that a test measures what it claims. Standard scores typically have a mean of 100 and standard deviation of 15, so 85–115 is within one SD. Percentile rank is not percent correct; grade equivalents overstate precision.
Data-based decisions
RTI/MTSS uses universal screening, tiered intervention, and progress monitoring. Curriculum-based measurement (oral reading fluency, maze, math computation) is brief, repeatable, and sensitive to growth. Set a baseline, draw an aimline to the goal, and adjust when several consecutive data points fall below the line-or when a trend is flat despite fidelity.
Traps
Confusing reliability with validity, treating one low score as a diagnosis, and ignoring the standard error of measurement. Accommodations (extended time, read-aloud) do not change the construct; modifications do. Also watch for alternate assessment questions that ignore eligibility criteria.
Worked example
Baseline 40 wpm, goal 70 wpm in 20 weeks: the aimline rises 1.5 wpm per week. Four consecutive points below the aimline signal a need to change the intervention, not wait.
Rule of thumb
Use multiple sources and repeated measures; a single score is a data point, not a decision.