Adaptive tests explained, for parents
Why your child says the test "got harder", what the score actually measures, and why half the questions are supposed to feel difficult.
Children in California meet adaptive testing in two main places: the state's CAASPP / Smarter Balanced assessments, and MAP Growth, which many districts use two or three times a year to track progress. Both work on the same principle, and both produce results that confuse parents for the same reason.
What "adaptive" actually means
A traditional test gives every child the same questions. An adaptive test chooses the next question based on the previous answer. Answer correctly and the next question is harder; miss one and the next is easier. The test keeps narrowing in until it finds the level where the child answers roughly half the questions correctly.
That last point is the one worth repeating to a child: a well-functioning adaptive test should feel hard about half the time. A child who finds every question easy has not been measured; a child who found most questions hard has not failed. Many children come home convinced they did badly purely because the test did its job.
Why the score is not a percentage
Because every child sees different questions, "number correct" is meaningless — twelve correct on hard questions is a very different result from twelve correct on easy ones. So adaptive tests report a score on a stable scale instead. MAP Growth uses the RIT scale; CAASPP reports a scale score and an achievement level.
The useful property is that the scale does not reset each year. A score of 200 means the same thing in Grade 3 as in Grade 6, which is what makes growth visible. A child moving from 195 to 207 has learned something real, even if their percentile barely moved — percentiles compare them to other children, while the scale measures what they can do.
Reading a report without over-reading it
- Look at growth before level. One score is a snapshot with an error range attached. Two scores six months apart tell you whether things are moving.
- Every score has a margin. Reports often show a range rather than a single number for good reason. A five-point difference between two sittings is usually noise, not change.
- The strand breakdown is where the useful information is. "Geometry below, Number and Operations above" tells you what to practise. The headline number does not.
- One bad morning explains a lot. Illness, a late night, or a child who clicked through to finish early all show up as a dip.
How to prepare, and how not to
Drilling test questions is largely wasted effort, because the test adjusts to whatever level the child is working at. What does help is fluency with the things every question depends on: number facts, fractions, and reading carefully enough to notice what is being asked.
What also helps, and is often overlooked, is familiarity with the experience: that you cannot go back, that you are not told whether an answer was right, and that difficulty rising is a good sign. A child who understands those three things walks in calmer.
Our adaptive practice test reproduces that experience in maths and English for grades 1 to 10 — difficulty adjusts as you answer, you are not told right or wrong until the end, and the result comes with a topic breakdown and the questions worth going over. The score it reports is our own estimate on a similar scale, not an official MAP or CAASPP score, and no school will accept it as one. It is there to show a child what the format feels like and to point at weak topics, nothing more.
Where to find the official practice
California publishes free practice and training tests for CAASPP, and NWEA publishes guidance for MAP Growth. If a real assessment is coming up, use those alongside anything else — they are the authentic format.