Which correlation coefficient assesses the consistency of measures within the same class or context and is used in multilevel models to measure dependency?

• A. Pearson correlation
• B. Intraclass correlation (ICC)
• C. Spearman correlation
• D. Canonica l correlation

Answer: (B)

The main idea here is measuring how consistent measurements are when they come from the same group or context, which reveals dependency among observations. Intraclass correlation (ICC) is designed exactly for this: it quantifies how similar measurements are within the same group and, in multilevel models, it gauges the degree of dependency or clustering in the data. A higher ICC means more of the total variation is due to differences between groups rather than within groups, indicating strong within-group agreement.

This is the best choice because Pearson correlation looks at linear association between two variables across all data, not how measurements cluster within groups. Spearman correlation is the rank-based counterpart of Pearson, still about association between two variables rather than within-group consistency. Canonical correlation examines the relationship between two sets of variables, not the reliability of measurements within a single group or context.