Dear All, I am unsuccessful in replicating the results from Field 2005 for ICC using ICC command in R, I would greatly appreciate your input in helping with the correct coding to get these results. Data are bellow as are the ICC results:
Table 1 Marks on eight essays by four lecturers
Essay Dr Field Dr Smith Dr Scrote Dr Death Mean S2 S2(k − 1)
1 62 58 63 64 61.75 6.92 20.75
2 63 60 68 65 64.00 11.33 34.00
3 65 61 72 65 65.75 20.92 62.75
4 68 64 58 61 62.75 18.25 54.75
5 69 65 54 59 61.75 43.58 130.75
6 71 67 65 50 63.25 84.25 252.75
7 78 66 67 50 65.25 132.92 398.75
8 75 73 75 45 67.00 216.00 648.00
Mean: 68.88 64.25 65.25 57.38 63.94 Total: 1602.50
my attempt:
# clear environment
knitr::opts_chunk$set(cache = TRUE)
setwd("C:/Users/..")
getwd()
library(data.table)
library(rio) # Easy way to read in files of varying formats.
library(psych)
dat1 <- import("Exp5_a.xlsx")
DT <- as.data.table(dat1) # convert to data.table
DT2 <- dcast(DT, text ~ rater , value.var = "score" ) # reshape data
DT2 <- DT2[, !c("rater", "text")] # drop unwanted variables
inter <- ICC(DT2)
DT4 <- DT[ , .(mean_q = mean(score)), by = c("rater", "text")]
DT5 <- dcast(DT4, text ~ rater, value.var = "mean_q")
xx <- 1:4
names(DT5) <- c("text", paste0("rater", xx))
DT5
DT6 <- DT5[, !c("text")]
ICC(DT6)
Kind regards,
Richard