irrICC (version 1.0)

icc3.inter.fn: Intraclass Correlation Coefficients (ICC) under the Mixed Factorial ANOVA model with Interaction.

Description

This function computes 2 ICC estimates ICC(3,1) and ICCa(3,1) as measures of inter-rater reliability and intra-rater reliability coefficients under Model 3. It is the the mixed factorial ANOVA model with interaction. Some subjects are expected to have multiple ratings and due to the assumed interaction effect.

Usage

icc3.inter.fn(dfra)

Arguments

dfra

This is a data frame containing 3 columns or more. The first column contains subject numbers (some duplicates are expected, as some subject are assumed to have assigned multiple ratings) and each of the remaining columns is associated with a particular rater and contains its numeric ratings.

Value

This function returns a list containing the following 12 values: 1. sig2s: the subject variance component. 2.sig2r: the rater variance component 3. sig2e: the error variance component. 4. sig2sr: the subject-rater interaction variance component. 5. icc2r: ICC as a measure of inter-rater relliability. 6. icc2a: ICC as a measure of intra-rater reliability. 7. n: the number of subjects. 8. r: the number of raters. 9. max.rep: the maximum number of ratings per subject. 10. min.rep: the minimum number of ratings per subjects. 11. M: the total number of ratings for all subjects and raters. 12. ov.mean: the overall mean rating.

References

Gwet, K.L. (2014): Handbook of Inter-Rater Reliability - 4th ed. - Equation #10.2.9 of chapter 10, page 279 (for the inter-rater reliability ICC(3,1)), and Equation #10.2.10 of chapter 10, page 279 (for the intra-rater reliability ICCa(3,1)), Advanced Analytics, LLC.

Examples

# NOT RUN {
#iccdata2 is a small dataset that comes with the package. Use it as follows:
library(irrICC)
iccdata2 #see what the iccdata2 dataset looks like
icc3.inter.fn(iccdata2[,2:6]) #Here, you must omit the first column
coeff <- icc3.inter.fn(iccdata2[,2:6])$icc2a #this gives you intra-rater reliability coefficient
coeff
# }