Assessment and Data-Driven Practice
Assessment and Data-Driven Practice tests how you select, read, and use data. Expect items on reliability and validity, score interpretation, screening, and turning numbers into program decisions.
Core rules
- Reliability is consistency (test-retest, internal consistency, inter-rater); validity is whether a score supports the intended inference (content, construct, criterion). A test can be reliable but invalid for a given use.
- Standard error of measurement reminds you that an observed score is a band, not a point; avoid high-stakes decisions from one score.
- Percentile rank means the percentage of the norm group scoring at or below that score. It is not percent correct, and percentile gaps are not equal units. Stanines split the normal curve into nine units; stanine 5 spans roughly the middle 20%.
- Norm-referenced vs. criterion-referenced: national comparison vs. mastery of specific skills.
- Data types: process (did we deliver it?), perception (survey), outcome (grades, attendance, discipline, graduation).
Use universal screening and progress monitoring, disaggregate by subgroup, find the root cause, write SMART goals, and report results.
Worked example
A student scores at the 40th percentile in reading. A colleague calls it 40% correct. Wrong. It means 40% of the national norm group scored at or below her score - near the middle, not failing. The team pairs this with work samples, attendance, and screening data before recommending intervention, then monitors progress every three weeks.
Traps
Equating percentile with percent correct; treating a small sample as a trend; confusing correlation with cause; letting one score drive placement; using an assessment outside your training or without required qualifications and consent; reporting subgroup data without protecting privacy.