A student can say 6 times 7 when asked in a quiet room. On the timed sheet he finishes 9 of 40 items. His teacher says "he knows his facts." The graph says something different. Both are true.
Fluency, knowing a fact quickly and accurately, is what a timed probe measures. This guide covers how to pick and give a math probe, how to score it, and how to turn the scores into an IEP baseline and goal.
Why computation
The Texas Education Agency's progress monitoring system describes CBMs as timed tests that are an important indicator of general proficiency, and says that for math the indicator is often computational fluency. It cites the National Mathematics Advisory Panel's 2008 finding that computational fluency predicts overall math performance from kindergarten through grade 6. The same document is careful to say fluency is only one strand of mathematical proficiency.
That caution matters for goals. A fluency probe tells you how fast and accurate a student is with computation. It does not tell you whether the student can reason through a word problem.
Single-skill or general
NCII's goal-setting guide draws a line that matters for facts.
| Single-skill probe | General outcome probe | |
|---|---|---|
| Example | 40 addition facts to 10 | A sheet that samples the year's computation skills |
| Tells you | Whether a skill is mastered | Growth across the year |
| NCII on goal-setting | Strategies in its guide should not be used | Benchmarks, national rates of improvement, and own-growth formulas apply |
If your student is working on one operation, a single-skill probe is a fair way to track mastery of it. But NCII says the goal-setting strategies in its guide are for general outcome measures, and that single-skill goals instead require statements about the number of objectives to be met by the end of the year.
Timing and scoring
Texas's system gives a concrete model. Students get a fixed time, and the time depends on grade.
| Grade | Time |
|---|---|
| K and 1 | Two 1-minute tests, one for each side of the page |
| 2 and 3 | 2 minutes in total |
| 4, 5 and 6 | 4 minutes in total |
The system has 20 forms per grade, each with 30 to 40 items, and every form follows the same item pattern so that scores compare. The tester reads a script, starts a timer at "Start," avoids answering questions about the items, and says "Stop." The score is the number of items answered correctly in the time, plotted on a graph.
NCII's math example is stated in digits per week, which points to a digits-correct score. Both ways of scoring are used. Choose one with your tool, and keep it for the whole goal.
A baseline for a probe
NCII describes three ways to get a baseline score, including the median of three probes. For a computation probe, say a student gets 14, 18, and 16 items right on three 2-minute probes. In order they are 14, 16, 18, and the median is 16.
Write the baseline with the unit and the conditions: "16 items correct in 2 minutes, median of three probes on Sept 8, 10, and 12, grade 3 form, quiet room." For the full method, see how to set an IEP baseline.
A goal for a probe
NCII's guide gives a template: when given the grade level and tool, the student will do the behavior at a level of proficiency within a time frame. A matching math goal, with invented numbers, reads: "When given a grade 3 computation probe for 2 minutes, Marco will answer 30 items correctly on three consecutive probes by the spring benchmark."
To check whether 30 is realistic, work out the weekly growth: from 16 to 30 is 14 items. With 28 weeks left, that is 0.5 items a week. Compare that with the rate of improvement your tool publishes. See rate of improvement explained. NCII's own math example shows how the formula works: 2 digits per week times 10 weeks plus a baseline of 30 digits gives a goal of 50.
Add accuracy
Items correct in a fixed time hides the difference between a student who attempts 20 and gets 18 right, and one who attempts 40 and gets 18 right. Record items attempted as well. Accuracy is correct divided by attempted: 18 of 20 is 90 percent, and 18 of 40 is 45 percent. The second student probably needs different teaching, perhaps for the facts themselves, and the first for speed.
This is practice guidance and not a rule from the sources. But recording both numbers costs seconds and changes what you teach next.
Charting facts you are teaching
For a student on single-skill goals, such as the 3s, 4s, and 6s multiplication facts, a simple mastery grid works. List each fact family down the side and dates across the top. Mark each date with the probe result, and write the criterion that counts as mastered, for example 36 of 40 in 2 minutes on two probes. This does not replace a general probe, but it shows the step-by-step progress that a single yearly score hides. Use it alongside, not instead.
Fluency is one strand
Texas says computational fluency is only one strand of mathematical proficiency. If a student's math need is broader than speed with facts, add a second measure. NCII's list of sample target skills includes number concepts and applications for grades 2 to 8, alongside computation. A pair of measures, one for fluency and one for concepts, gives the team a fuller view and keeps the goal from narrowing to whatever the timed sheet rewards.
Giving it the same way
- Same time limit every week.
- Same form type. Texas keeps the item pattern the same across forms.
- No hints. Texas says avoid feedback.
- Same scoring unit. Digits or items, not both.
- Note conditions. A noisy room or a missed breakfast is worth one line.
Math probe errors
- Using a single-skill probe with benchmark formulas. NCII says not to.
- Changing the time limit to help the student. It changes the measure.
- Counting partial credit inconsistently. Decide before the first probe.
- Teaching only the probe items. Texas uses alternate forms so that scores reflect skill, not memory of the sheet.
See also curriculum-based measurement basics and the glossary.
Other publishers, other rules
The sources do not set a standard number of probes or a required digit rule. Your tool's manual and your IEP team decide. Texas's system is one example, and other publishers use different times and forms.