Strategic StudentIs this difference real?
You saw a number move. Maybe chronic absence dropped five points, or one group is outscoring another. Before you call it progress or a problem, check whether the difference is big enough to trust, or small enough to be chance. No PII. No login. Everything runs in your browser.
Enter the two numbers you're comparing
No PIIYour read will appear here
Enter a rate and a group size for each side above. We'll tell you whether the difference is statistically real or likely just noise.
How big the difference really is
The shaded band is the range the true difference most likely falls in. If it touches the center line, we can't rule out no difference.
Keep in mind
How this works
This tool runs a two-proportion z-test, the standard way to ask whether two rates differ by more than chance. It compares the gap between your two rates to how much each rate could wobble given the number of students behind it. When one group has very few students on one side of the question (an expected count under 5), it switches to Fisher's exact test, which stays accurate where the z-test does not.
The 95% confidence interval for the true difference uses the Newcombe hybrid score method. It holds up with small groups and with rates at or near 0% or 100%, exactly where the simple textbook interval collapses to a single point.
A result is flagged "likely real" when it would happen by chance less than 5% of the time (p < 0.05), the common bar in education research. Small groups make almost any difference look like it could be chance, which is exactly why group size matters as much as the gap.
Statistical significance isn't the same as importance. A tiny gap can be "real" with huge groups, and a meaningful gap can be "uncertain" with small ones. Always read it alongside what you know about your students. Sources: Newcombe, R. G. (1998). Two-sided confidence intervals for the single proportion: comparison of seven methods. Statistics in Medicine, 17, 857-872, and Newcombe, R. G. (1998). Interval estimation for the difference between independent proportions: comparison of eleven methods. Statistics in Medicine, 17, 873-890.
A significance test tells you a difference is probably not a fluke. It'll never tell you it matters for a student. Use it to decide what's worth a closer look, then bring your judgment.