The reliable change index, or RCI, tells you whether a person’s change in score is larger than measurement error. The reliable change index calculator takes the scores before and after, the standard deviation of the scale and its reliability, and gives the standard error of measurement, the RCI, the p-value and the smallest reliable change.
How the calculation works
- Compute the standard error of measurement: SEM = s × √(1 − r), where s is the standard deviation of the scale in a reference group and r its test-retest reliability.
- The standard error of the difference between two scores of the same person is Sdiff = √2 × SEM, because both scores carry measurement error.
- The reliable change index of Jacobson and Truax is RCI = (score after − score before) ÷ Sdiff.
- An absolute RCI above 1.96 means the change is unlikely, with 95 % confidence, to come from measurement error alone. The smallest reliable change is 1.96 × Sdiff.
Example
A person scores 62 before treatment and 48 after, on a scale with a standard deviation of 10 and a reliability of 0.85. SEM = 10 × √0.15 = 3.873, Sdiff = 5.477, and RCI = −14 ÷ 5.477 = −2.556. The p-value is 0.0106 and the smallest reliable change is 10.74 points, so the 14-point drop is a reliable change.
Why it matters
Scores move from one occasion to the next even when nothing has changed. Without a yardstick it is easy to over-read a small difference as progress or decline. The RCI gives a clear threshold for when a change is more than noise.
Good to know
The index treats the reliability and the standard deviation as known values from a good reference sample. It does not correct for practice effects or regression to the mean, which can make a second score drift even for a stable person. Read it as a screening threshold and not as proof of change.
Related tools
The standard error calculator covers the error of a sample mean, and the paired t-test calculator tests change at the group level.