Brandon A. BakiData literacyFor teachers and leaders · one page
Five traps when reading assessment data
Data can show you where to act – or send you the wrong way. Five mistakes even experienced readers make, shown with MAP Growth examples, and what to do instead.
-
The average hides the spread
Two classes can share an average of 212 and need very different teaching: one tightly grouped, one running from struggling to far ahead.
Instead Look at the range, the quartiles and who sits at each end first.
-
Small groups swing
In a group of six, one student is 17 percentage points. A jump from 33% to 67% can be two students having a good day.
Instead Put the number of students beside every percentage; read small groups over several terms.
-
A percentile is not a score
Percentiles bunch at the ends. Ten scale points lift a student in the middle 26 percentile points; a student near the top, only 7.
Instead Use the scale score (RIT) for growth and gaps; percentiles to compare with national norms.
-
High scores can hide low growth
Achievement says where a student is; growth says how far they moved. A high achiever who has stalled needs attention too.
Instead Read achievement and growth together: a quadrant chart shows both at once.
-
Not every test counts the same
A test finished in nine minutes, a retest, a noisy room: the score may say more about the conditions than the learning.
Instead Check duration and flags first; leave flagged tests out of averages, and say so.
Before you share a number, ask
- How many students?A percentage needs its n.
- Compared with what?Last term, the grade, the norm.
- Measured how?Which test, what conditions.
- What else explains it?Absence, a new class, a retest.