Which statistical measure is commonly used to assess the degree of agreement or consistency of measurements made by the same raters or instruments within a group?

• A. Pearson correlation
• B. Spearman correlation
• C. Intraclass correlation (ICC)
• D. Kendall's tau

Answer: (C)

The idea being tested is how reliable measurements are when the same raters or instruments measure the same subjects, focusing on how much the measurements agree with each other. Intraclass correlation is built for this purpose: it looks at the variability of scores within subjects across raters or occasions and compares it to the variability between subjects. When most of the variation comes from true differences between subjects rather than from measurement error or rater inconsistency, the intraclass correlation is high, indicating strong agreement or consistency.

This statistic is different from Pearson, Spearman, or Kendall correlations, which assess association or monotonic relationships between variables rather than actual agreement in the measured values. For instance, two raters could move their scores up and down together (high correlation) but still disagree by a consistent amount, which would yield a lower ICC. Intraclass correlation can handle multiple raters and repeated measurements and can be computed in forms that reflect whether you care about absolute agreement or just consistency.

So, when you want to know how similar the measurements are within a group of raters or instruments measuring the same subjects, intraclass correlation is the appropriate choice.