Clarifying Discrepancies in Responsiveness Between Reliable Change Indices.

Area of Science:

• Neuropsychology
• Psychometrics

Background:

• Reliable Change Indices (RCIs) are used to detect statistically significant individual changes in repeated neuropsychological assessments.
• Guidance on selecting the appropriate RCI model and understanding its implications is limited.

Purpose of the Study:

• To systematically evaluate key parameters influencing different RCI models using existing test-retest norms.
• To provide guidance on RCI model selection for neuropsychological assessment.

Main Methods:

• Utilized normative test-retest data for Wechsler Memory Scale-IV subtests, considering differential practice effects.
• Manipulated individual relative position to the normative mean to assess RCI responsiveness.

Main Results:

• RCI responsiveness varied based on baseline scores relative to the normative mean and retest variance.
• RCI McSweeny was most responsive for scores below the mean; RCI Chelune and Maassen differed in responsiveness for scores above the mean based on variance.
• Responsiveness patterns reversed for positive change and could be influenced by excellent test-retest reliability with greater variability.

Conclusions:

• RCI models generally agree when baseline scores are near the normative mean and variability is equal.
• No single RCI model is universally superior; model choice impacts change interpretation.
• A more informed selection of RCI models is now possible, with a preference for regression-based models noted.