Repeated Measures Correlation

Area of Science:

• Psychology
• Statistics
• Behavioral Science

Background:

• Simple regression/correlation can yield biased results with non-independent or aggregated data.
• Violations of independence and differing between- vs. within-participant patterns are common issues.
• Repeated measures correlation (rmcorr) addresses these limitations for paired measures.

Purpose of the Study:

• To introduce and explain the repeated measures correlation (rmcorr) technique.
• To provide guidance on its assumptions, equations, visualization, and statistical power.
• To demonstrate rmcorr's application using an R package and example datasets.

Main Methods:

• Rmcorr analyzes paired measures from multiple individuals across two or more occasions.
• It estimates the common within-individual regression slope, representing the shared association.
• The study introduces an R package (rmcorr) for inferential statistics and visualization.

Main Results:

• Rmcorr does not violate the independence of observations assumption.
• It offers greater statistical power by avoiding data averaging or aggregation.
• The technique effectively illustrates intra-individual and inter-individual research questions.

Conclusions:

• Rmcorr is a robust statistical technique for analyzing common linear associations in paired repeated measures data.
• It provides a powerful alternative to simple correlation and aggregation methods.
• The R package facilitates reproducible research and accessible application of rmcorr.