Reliable Change Index (reliable change for each person)

In brief

What for: The Reliable Change Index (RCI) shows for each person whether the score changed between two measurements more than the measurement error of the test would lead us to expect (Jacobson & Truax, 1991). It is under "More options" in the question "Has something changed?". RCI = (after − before) / S_diff with S_diff = √2 · SD₁ · √(1 − r); SD₁ is the standard deviation of the before values. At |RCI| > 1.96 the change counts as reliable.

Required entries

• Reliability r: the retest reliability is the proper choice, usually from the test manual (the correlation of two measurements in unchanged people). Cronbach alpha from the data is only a substitute: it measures internal consistency, not stability over time, and is usually higher; changes then count as reliable sooner. The source of the reliability is stated in the result.
• "High values are better" or "worse": this decides whether an increase counts as an improvement or a worsening. Statable does not guess it.

Reading the result

• The sentence states how many people reliably improved, reliably worsened or did not change reliably. Exactly |RCI| = 1.96 does not count as reliable.
• The table shows for each person (row of the data) before, after, change, RCI and classification.

Limits

• Reliable only means larger than the measurement error. Whether the change is meaningful for the person (clinical significance) is not shown by the RCI; Statable does not calculate a cut-off.
• The RCI assesses individual people. It does not show whether an intervention works: that needs a comparison group. Extreme before values change anyway through regression to the mean.

Distinction from the typical error

The typical error describes the precision of the measurement itself, from two measurements of the same people without an intervention, and needs no reliability. The RCI judges whether a single person has changed reliably. More in the article "Measurement precision: typical error".

Source

Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59(1), 12-19.