coefRi: Calculate Interclass Correlation Coefficient (ICC)

In stufield/stuRpkg: Stu's Utilities Package

Calculate Interclass Correlation Coefficient (ICC)

Description

The Intraclass correlation coefficient (r_i; aka ICC) can be used to estimate the repeatability of a method. The value 0 -> 1. Depending on how the groups are set up, you want all your variation to be among groups (individuals), not within groups (repeats) so you want this value to be high if individuals are your groups, and low if your repeated measurements are the groups. When ICC is high, it means most of the variation is between treatment groups.

Usage

coefRi(x, groups)

Arguments

x	A matrix or data frame containing the raw data of the various treatments you are testing in the ANOVA.
groups	A vector of the factor groupings for x.

Value

A list containing:

model	Resulting ANOVA table
ICC	The intraclass correlation coefficient, the measure of similarity among individuals within a treatment group relative to the differences found among groups

Author(s)

Stu Field

References

Sokal & Rohlf (Biometry; 3rd ed.) 210-214. Sokal & Rohlf (Biometry; 2rd ed.) 211-216.

See Also

aov()

Examples

head(Ri_data)    # internal data
x <- as.vector(as.matrix(Ri_data))[ !is.na(as.vector(as.matrix(Ri_data))) ]
coefRi(x, groups = rep(names(Ri_data), c(8, 10, 13, 6)))

stufield/stuRpkg documentation built on April 2, 2022, 2:05 p.m.