Which test is the parametric method for reliability?

Which test is the parametric method for reliability?

• A. ICC
• B. Kappa
• C. Pearson
• D. Spearman

Answer: (A)

Reliability of measurements across raters or instruments when the data are on a continuous scale is best assessed with the intraclass correlation coefficient. This statistic comes from an ANOVA-based partitioning of variance into components attributable to subjects, raters, and random error. By comparing the amount of true variation between subjects to the total observed variation, ICC tells you how consistently measurements reflect the same underlying trait. Its foundation in variance components and distributional assumptions makes it a parametric method for reliability. Kappa is used for categorical data to assess agreement beyond chance, not for continuous scores. Pearson correlation measures linear association between two continuous variables but does not separate measurement error or rater effects, so it isn’t a reliability measure across raters. Spearman is similar to Pearson but uses ranks and is nonparametric; it also does not quantify reliability across raters in the variance-components sense. Therefore, the parametric method for reliability among continuous measurements is the intraclass correlation coefficient.

Reliability of measurements across raters or instruments when the data are on a continuous scale is best assessed with the intraclass correlation coefficient. This statistic comes from an ANOVA-based partitioning of variance into components attributable to subjects, raters, and random error. By comparing the amount of true variation between subjects to the total observed variation, ICC tells you how consistently measurements reflect the same underlying trait. Its foundation in variance components and distributional assumptions makes it a parametric method for reliability.

Kappa is used for categorical data to assess agreement beyond chance, not for continuous scores. Pearson correlation measures linear association between two continuous variables but does not separate measurement error or rater effects, so it isn’t a reliability measure across raters. Spearman is similar to Pearson but uses ranks and is nonparametric; it also does not quantify reliability across raters in the variance-components sense. Therefore, the parametric method for reliability among continuous measurements is the intraclass correlation coefficient.